睡眠与分子：生物化学与化学生物学的贡献｜老夫之乐

（2024年12月30日英文文章上线）

用生物化学、化学生物学和遗传学揭示睡眠调控的必要分子：钙调磷酸酶

余建军^1,2,*,刘慧洁^1,2,*,高瑞^1,*,王涛^1,2,*,李成钢^1,2,*,刘玉祥^1,2,杨璐^1,2,徐颖³,崔云凤²,贾辰熙³,黄娟¹,陈鹏¹,饶毅^1,2,4,# 

¹Laboratory of Neurochemical Biology, Peking-Tsinghua Center for Life Sciences,Peking-Tsinghua-NIBS (PTN) Graduate Program, School of Life Sciences, Peking University, Beijing, China; Chinese Institute for Brain Research, Beijing (CIBR); Department of Chemical Biology, College of Chemistry and Chemical Engineering; School of Pharmaceutical Sciences, PKU-IDG/McGovern Institute for Brain Research, Peking University, Beijing, China.

²Chinese Institutes for Medical Research, Beijing (CIMR), Capital Medical University, Beijing, China.

³National Center for Protein Sciences Phoenix, Beijing, China.

⁴Lead contact.

摘要

传统上，研究睡眠的机理依赖于电生理学和遗传学。由于睡眠只能通过行为观察和物理学手段在整体动物上测量，因此没有通过生化和化学生物学方法来找到调控睡眠分子的先例。此前发现影响睡眠的蛋白激酶SIK3，其磷酸化位点对于功能很重要，我们以此为靶点，用生物化学和化学生物学的方法作为这项研究的起点。我们发现在小鼠的SIK3进行点突变后（将469位的苏氨酸变为丙氨酸，T469A），突变小鼠的睡眠增加。我们使用了生化纯化和光交联两个方法，却殊途同归地发现，钙调磷酸酶（简称CaN）无论在体外和体内还是体内，都能去除SIK3的T469（和S551位）的磷酸基团，但却不影响SIK3在T221位点的磷酸化。降低脑内CaN调节亚基的的基因表达量后，小鼠每昼夜睡眠时间减少5个小时以上，变化幅度超过所有已知的小鼠遗传突变体。我们发现了CaN在小鼠睡眠中的关键生理作用，并开创了用生化纯化和化学生物学揭示睡眠分子机理的有效途径。

意义

睡眠的重要性不言而喻，但其分子和细胞机制仍然是个谜。包括我们在内的科学家已用遗传学方法在果蝇和小鼠中寻找参与睡眠调控的基因。但是在本研究中，我们率先应用生化和化学生物学方法来揭示睡眠调节的机理。我们发现钙磷酶（CaN）控制SIK3 去磷酸化有位点特异性：可以作用于两个调节位点，但不能作用于酶活性所需的位点。用分子生物学方法降低CaN 会降低睡眠5 个多小时，这是在遗传突变小鼠身上观察到的最显著睡眠表型。这项研究揭示了睡眠调节中的蛋白磷酸酶-蛋白激酶途径，彰显了生化纯化和化学生物学方法作为研究大脑功能有效技术的价值。

本研究利用培养细胞进行蛋白质分离纯化，脑组织光交联，质谱分析脑内磷酸化蛋白组，发现磷酸酶对同一个蛋白质上不同位点的高度选择性去磷酸化，在体外用细菌表达的蛋白质重组磷酸酶活性，同时用了4种基因修饰小鼠、两种病毒介导的基因敲低的小鼠，密切结合分子生物学、生物化学、遗传学和脑电图记录而阐明基因与睡眠的关系。

引言

睡眠是动物的重要生理过程¹，受昼夜节律和稳态过程的调节^2,3。事实证明，在果蝇、小鼠、狗和人类身上采用遗传学方法能有效发现调节睡眠的基因^4-10。例如，发现了食欲素及其受体在维持觉醒方面的作用^11-13，以及盐诱导激酶3（SIK3）在调节睡眠方面的作用¹⁴。

由于睡眠只能在动物身上测量，而不能在分子上测量，因此我们最初和本领域的其他研究人员一样，都是依靠电生理学和遗传学来发现果蝇和小鼠睡眠分子和机理。

在评估了这些策略的优势和局限之后，我们决定尝试同时将经典的生物化学纯化方法和现代的化学生物学途径用于研究睡眠调控机制。我们的第一个目标底物是 SIK3¹⁴，Yanagisawa组^14,20,21和我们自己^18,19发现其特异位点磷酸化对睡眠调控非常重要。此前我们发现SIK3的特异磷酸化位点可指示睡眠需求并参与睡眠调节¹⁹，因此我们使用SIK3蛋白作为底物，并对其调节因子进行生化鉴定，然后再在体内测试这些体外发现的磷酸化调节因子是否也能在体内调控睡眠。

蛋白质磷酸化水平在不同的睡眠-觉醒相关状态中有所不同^22-25。哺乳动物的睡眠被认为涉及蛋白激酶 A（PKA）^26-30、细胞外信号调节激酶（ERK）^31-33、单磷酸腺苷（AMP）激活蛋白激酶（AMPK）^34-36、钙（Ca²⁺）/钙调蛋白（CaM）激酶 II（CaMKII）a和b^37-39、c-Jun N 端激酶（JNK）40、SIK3、1 和 2^14、19-21 以及肝脏激酶 B1（LKB1）^18、41-43。CaMK2b基因敲低的小鼠被报道每24小时的睡眠时间减少了120分钟以上³⁸ ，超过了其他已知的基因突变小鼠^{9,12-14,18-21,27,30,33,39,44-53}。

蛋白磷酸酶（PPases）是否参与哺乳动物睡眠知之甚少。1970 年代发现一种蛋白质^54-56 被命名为钙调磷酸酶（也称钙神经蛋白，calcineurin, CaN、PP2或PPP3）⁵⁷ ，到1980 年代确定其作为磷酸酶的功能^58,59。CaN是唯一由钙离子（Ca²⁺）和钙调素（CaM）激活的磷酸酶，由二聚体组成，具有催化亚基A三种之一（PPP3CA、PPP3CB或PPP3CC）和调节亚基B两种之一（PPP3R1或PPP3R2）^56,60。PPP3CA、PPP3CB 和PPP3R1 表达的组织器官较为广泛，而PPP3CC和PPP3R2则在睾丸中特异表达^61-65。在脑中，PPP3CA 是含量最高的催化亚基，PPP3R1 是含量最高的调节亚基。

本项工作中，我们在确定了SIK3中469位点苏氨酸（T469）磷酸化的重要性后，再开始寻找 SIK3的磷酸酶。虽然小鼠SIK1和SIK2中与SIK3丝氨酸（S）551 同等位点的缺失会导致与SIK3 S551A相同的功能增益（gain of function, GOF）表型²¹，但我们发现Sik1或Sik2基因的缺失不影响小鼠的睡眠¹⁹。因此，我们重点研究Sik3。PKA磷酸化SIK3 T469和S551位点后促进SIK3与14-3-3蛋白的相互作用⁶⁶。生化上，14-3-3蛋白抑制SIK3活性，而T469或S551 的缺失增加了 SIK3 的信号转导⁶⁶。T469磷酸化的体内生理学功能意义尚不清楚。在此，我们将小鼠SIK3的T469突变为丙氨酸（A）（T469A），发现它导致小鼠睡眠增加。

我们采用了两种途径寻找去磷酸化SIK3的T469和S551的磷酸酶：从人胚胎肾脏（HEK）293T细胞中用经典生物化学纯化T469和S551的磷酸酶，以及在小鼠脑中用化学生物学光交联寻找与SIK3相互作用的蛋白。都找到了CaN的催化亚基PPP3CA。进一步体外生物化学重组实验表明，在 Ca²⁺、CaM和PPP3R1的存在下，PPP3CA可去除T469和S551的磷酸基团。PPP3CA 不能去除T221的磷酸基团，后者的磷酸化增强SIK3的激酶活性，并促进睡眠，且其水平可以反应睡眠需求¹⁹。在体外培养的HEK293T细胞中，敲除PPP3CA、PPP3CB 或 PPP3R1，可以抑制钙离子诱导的 T469和S551（而非 T221）去磷酸化。在鼠脑通过分子生物学敲除PPP3CA或PPP3R1时，鼠脑中 T469 和 S551 的磷酸化水平增加，T221的磷酸化不受影响。敲除PPP3CA 基因在 24 小时内减少睡眠约 3小时（187.6±13.0分钟）。敲除PPP3R1基因减少5小时以上（349.3±21.5分钟），超过了所有已知小鼠遗传突变体的睡眠变化。从睡眠变化的程度来看，CaN是迄今为止发现的最重要的睡眠调节因子。我们成功地利用生化纯化和化学生物学发现了重要的睡眠调控分子，显示了遗传学外，还有生物化学和化学生物学对睡眠研究有巨大潜力。

结果

T469在调节小鼠睡眠中的功能意义

我们首先构建了一株携带 T469A 点突变的小鼠品系（图 1、S1-S4、表 S1 和 S2）。睡眠通常是在雄性小鼠中测量。由于基因型为Sik3^T469A/T469A的点突变纯合体雄性小鼠胚胎致死，我们比较了两种基因型（Sik3^+/+野生型纯合体和Sik3^T469A/+点突变杂合体）的雄性小鼠睡眠表型。我们还研究了所有三种基因型（Sik3^+/+和Sik3^T469A/+和Sik3^T469A/T469A点突变纯合体）雌性小鼠的睡眠表型（图S3和S4，表S1和S2）。雄、雌小鼠的睡眠表型在趋势上相似，雌性Sik3^T469A/T469A点突变纯合体的表型强于雄性 Sik3^T469A/+点突变杂合体，在总睡眠和非快速眼动睡眠（NREMS）方面分别相差25.2分钟和 37.1分钟（表S2，雌性Sik3^T469A/T469A和Sik3^+/+之间的睡眠增量与雄性 Sik3^T469A/+和Sik3^+/+之间的睡眠增量）。

图1

代表性脑电图（EEG）和肌电图（EMG）见图 1L，典型的脑电热图见图 S2E。与 Sik3^+/+ 相比，Sik3^T469A/+雄性24 小时内总睡眠时长以及非快动眼睡眠（NREMS ）持续时间增加，尤其是在暗期（图1A、1B、1C，表S2）。NREMS发生次数和单次持续时间均无明显变化（图1D和1E）。快速眼动睡眠（REMS）在Sik3^T469A/+和Sik3^+/+雄性之间无显著差异（图S1 E-H，表S2）。脑电图的功率谱在δ波段增加、α波段减少（图1F和1G）。通过NREMS δ波功率密度测量的睡眠需求在Sik3^T469A/+雄性显著增加^2、67-70（图1H）。

Sik3^T469A/T469A雌鼠在暗期的NREMS时长显著增加（图S3A和 S3B，表S1和S2），原因是NREMS发生次数增加（图S3C），但单次持续时间无变化（图S3D）。REMS时长在暗期没有差异，但在亮期有所减少（图S3E和S3F，表S1和S2），原因是REMS 发生次数减少（图S3G）、单次持续时间无变化（图S3H）。从NREMS到REMS的转换几率减少，而其他状态间转换几率没有变化（图S3 M-P）。与Sik3^+/+雌性相比，Sik3^T469A/T469A和Sik3^T469A/+雌性的 NREMS δ波功率密度都有所增加（图S4E）。

Sik3^T469A/+雄性小鼠的睡眠表型与另一实验室报道的Sik3^S551A/+不同²⁰。我们随后构建了Sik3^S551A/+突变小鼠。我们的Sik3^S551A/+雄性小鼠表现出与 Sik3^T469A/+雄性小鼠高度相似的睡眠表型，即在暗期NREMS 持续时间和δ功率密度增加（图S5A-B、S6E、表S3和S4），但清醒（Wake）至 NREMS的过渡几率和睡眠剥夺（sleep deprivation, SD）后REMS 反弹降低（图S5M、图S6G）。在生化上，我们发现T469A 或 S551A点突变会降低另一个位点的磷酸化水平（图S6M和S6N）。

我们也构建了Sik3^T469E/+点突变小鼠，但没有观察到任何睡眠表型（图S7和S8）。

SIK3的T469和S551磷酸酶的生化纯化

虽然 SIK3 功能丧失性突变（loss of function, LOF）和GOF小鼠突变体只表现有限的睡眠表型，但 SIK3 的上游或下游可能存在更重要的调节因子。为研究这种可能性，我们从 HEK293T 细胞中纯化了其上游磷酸酶，使用SIK3 在 T469 和 S551 位点的去磷酸化程度作为磷酸酶的活性指标。我们从 HEK293T 细胞中纯化能使 SIK3 在T469 和 S551 位点去磷酸化的磷酸酶（图2）。

图2

我们先用大肠杆菌（E. coli）合成并纯化的重组 SIK3 蛋白证实了抗SIK3 T469 和 S551磷酸化形式抗体的特异性（图S9A）。

随后，我们裂解了 HEK293T 细胞，并将含500 mg总蛋白的细胞裂解液（浓度为 10 mg/ml）依次在 Q HP、Blue HP、SP HP、heparin HP、HAP HP和Superdex 200 色谱柱上分馏（图2A）。每一步都从每个馏分中抽取相等组分检测对T469 和 S551的磷酸酶活性。将一个步骤中的活性馏分合并，并在下一个色谱柱上进一步分馏。因此，Q HP 柱的10至 13号馏分、Blue HP 柱的流过（FL）馏分、SP HP 柱的FL馏分、heparin HP柱的FL馏分以及 HAP HP 柱的 5 至 7 号馏分被分别装载到下一个色谱柱上（图2B-G）。

在纯化过程中，我们观察到针对T469和S551的磷酸酶活性高度相关（图2B-F），而且在每一步纯化之后，整体磷酸酶活性都有所提高（图S9B）。

将来自Superdex 200柱的10至18号馏分通过银染检测蛋白含量（图2H）。其中馏分15含有最强的磷酸酶活性，从中切取一条箭头所示的条带进行质谱（MS）分析，最后我们检测到四种磷酸酶：PPP3CA、PPP3CB、PPP3CC 和 PPP5C（图2I）。

发现PPP3CA是小鼠脑内SIK3的相互作用蛋白

为研究小鼠大脑中的内源蛋白相互作用，我们（陈鹏和高瑞）发明了光交联方法来寻找与SIK3相互作用的内源蛋白。与神经元培养物或细胞裂解液相比，脑切片能更好地保存原位蛋白质互作网络。在目前研究蛋白质-蛋白质相互作用（protein-protein interactions, PPIs）的方法中，通过光照共价捕获相互作用蛋白的光交联策略因其良好的时间分辨率和低于化学交联的细胞毒性而被认为是在活体系统中更受欢迎的方法。然而，将光交联试剂加入感兴趣的蛋白质通常依赖于遗传密码扩增策略，即设计相应的tRNA合成酶以插入可光激活的非天然氨基酸⁷¹，或通过代谢过程插入含有氨基酸类似物的可光激活分子^72-74，而这两种方法都很难在组织中实现。此外，用于光交联的传统分子（如重氮啶和芳基叠氮化物）通常对紫外线照射敏感⁷⁵，而紫外线具有高能光毒性，组织穿透力弱⁷⁶。因此，在组织样本中原位捕获 PPI 的光交联策略还鲜有报道。

基于之前的工作⁷⁷，我们（陈鹏和高瑞）开发了一种光催化化学交联（photocatalytic chemical crosslinking，PhotoCAX）策略，用于捕获小鼠脑片中的 PPIs（图 3A）。我们选择eosin-Y作为光催化剂，1,6-diaminohexane作为连接剂。每种化合物都具有良好的溶解性，尤其是eosin-Y，被广泛用于组织染色。eosin-Y 的最大吸收波长为517 nm，在绿光照射下会产生单线态氧（¹O₂）⁷⁸，单线态氧会激活某些氨基酸残基（如酪氨酸）的侧链，形成亲电中间体，从而与胺弹头形成共价连接⁷⁹。我们的方法不需要转染或基因修饰就能将光催化剂或交联剂加入样品中进行 PPI 捕获。

我们随后构建了SIK3-3xHA小鼠，其中 SIK3 蛋白的羧基末端标记了血凝素 (HA) 抗原决定簇⁸⁰，以便我们能在光交联后检测与 SIK3-HA 相关的蛋白质。

我们在新鲜制备的 SIK3-HA 小鼠脑片中应用了 PhotoCAX 策略。样品中加入eosin-Y 和 1,6-diaminohexane，然后用绿色 LED（GL）照射。通过免疫印迹分析验证了SIK3相互作用组，其中+GL组出现了明显的交联条带，表明探针被有效激活（图3B）。

图3

在交联组和非交联组中，SIK3 蛋白都能被欧联抗-HA抗体的磁珠所富集，交联产生的相互作用组也能同时被拉下。为进一步分析蛋白质，我们在制备 MS 样品时切除了分子量大于 SIK3 的条带。在胰蛋白酶消化后，采用二甲基标记法对两组蛋白质进行定量研究。如Venn图（图3C）所示，81 个蛋白质在三个独立实验中重复出现，且在 +GL 组中富集程度较高（富集倍数 +GL/-GL > 4, log2(+GL/-GL) > 2）。通过火山图进一步分析富集结果（图3D），其中只有 p < 0.05 且3个重复中平均富集倍数超过4的蛋白质才被认为是 SIK3 的重要互作候选蛋白。作为阳性对照，我们检测到了一些已知的 SIK3 相互蛋白，如 14-3-3蛋白。PPP3CA 及其相应的调控亚基 PPP3R1也被富集（图3D）。

为证实SIK3和PPP3CA在鼠脑中的相互作用，我们在免疫沉淀实验中使用抗体富集SIK3，发现沉淀物中既有SIK3也有PPP3CA（图 3E）。同样，使用抗PPP3CA抗体处理鼠脑裂解液时，沉淀物中也发现了PPP3CA 和SIK3（图3F）。将FLAG标记的SIK3和HA标记的PPP3CA导入HEK293T 细胞过表达时，SIK3和PPP3CA的相互作用可在HEK293T细胞中再现（图 3G, 3H）。

体外PPP3CA对T469和S551而非T221的去磷酸化作用

我们通过免疫沉淀从HEK293T细胞中富集过表达的被FLAG 标记的 SIK3，这种SIK3在T221、T469和S551位点都已经被磷酸化。我们在HEK293T细胞表达和富集每个候选磷酸酶（PPP3CA、PPP3CB、PPP3CC、PPP5C和PPP3R1）。在PPP3R1存在时，SIK3的T469和S551可被PPP3CA、PPP3CB和PPP3CC去磷酸化。而PPP3CA、PPP3CB和PPP3CC不能去磷酸化T221。在相同条件下，PPP5C不能去磷酸化SIK3的T469、S551 或T221（图4A），但它能使Tau蛋白的S396位点去磷酸化（图 S9C），表明PPP5C具有催化活性。

图4

由于PPP3CA和PPP3R1在脑中含量丰富，而PPP3CC和PPP3R2在睾丸中含量丰富或特异表达^61-65,81，因此我们重点研究PPP3CA和PPP3R1。

在PPP3R1存在时，从HEK293T细胞富集纯化的PPP3CA可去磷酸化MEK1的S217，但不能去磷酸化AKT1的T308、PDK1的S241、MARK1的T215、GSK3β的S9、HDAC4的S246和S632、JNK1的T183/Y185以及 ERK2的T185/Y187（图4B-E）。因此，广为流传的所谓磷酸酶具有 “非特异性”对PPP3CA来说并不成立。

为排除从HEK293T细胞中富集纯化的酶含有与磷酸酶相关其他蛋白质的可能性，我们在大肠杆菌中表达了CaM、PPP3CA和PPP3R1。我们发现，体外重组SIK3的T469和S551去磷酸化活性需要CaM、PPP3CA、PPP3R1和Ca²⁺ 同时存在（图4F）。

在HEK293T细胞中，PPP3CA 和 PPP3R1对SIK3 T469 和 S551（而非 T221）的去磷酸化作用

Ionomycin是一种钙离子载体，可使Ca²⁺流入细胞。当对HEK293T细胞施用ionomycin时，ionomycin以剂量（图4G-H）和时间（图4I-J）依赖的方式使SIK3 T469和S551（而不是 T221）去磷酸化。

PPP3CA/PPP3R1的过表达可以增强ionomycin诱导的T469和S551的去磷酸化（图4K-L）。过表达无活性的PPP3CA形式PPP3CA-H151A（组氨酸或 H 突变为 A）不影响T469或S551的磷酸化水平（图4M），但降低ionomycin的去磷酸化活性，表明该 LOF 突变体具有显性负效应⁸²。

为研究ionomycin诱导的去磷酸化对CaN的依赖性，我们用sgRNAs 构建了 PPP3CA、PPP3CB 和 PPP3R1 的单基因或双基因敲除HEK293T细胞系。HEK293T细胞中PPP3CA或PPP3CB 单基因敲除的表型不稳定。然而，在HEK293T细胞中，PPP3CA 和 PPP3CB 双基因敲除或 PPP3R1 单基因敲除的表型稳定。PPP3R1 基因敲除或PPP3CA 和 PPP3CB 双基因敲除抑制ionomycin诱导的T469和S551（但不是 T221）的去磷酸化（图4N-P）。

PPP3CA和PPP3R1参与小鼠大脑中SIK3 T469和S551（而非 T221）的去磷酸化

为研究它们在体内的作用，我们使用 CRISPR-Cas9 策略在小鼠脑中靶向敲除PPP3CA或PPP3R1（图S10）。对照组有两种：WT^Ctrl是野生型（WT）小鼠，注射了靶向PPP3CA或PPP3R1的sgRNA表达病毒；eGFP^Ctr是表达Cas9的小鼠（Rosa^Cas9/+），注射了靶向增强型绿色荧光蛋白（eGFP）的sgRNA表达病毒；PPP3CA基因敲除（PPP3CA^KD）小鼠为Rosa^Cas9/+小鼠注射AAV2/PHP.eB-CMV-mScarlet-PPP3CA-sgRNA-WPRE病毒而产生；PPP3R1^KD小鼠为Rosa^Cas9/+小鼠注射AAV2/PHP.eB-CMV-mScarlet-PPP3R1-sgRNA-WPRE病毒而产生。

图5

免疫印迹分析表明，靶向PPP3CA或PPP3R1的sgRNA降低了其蛋白水平（图5）。PPP3CA被靶向时，PPP3R1蛋白也会减少（图5A）。在 PPP3R1^KD 小鼠中，PPP3R1、PPP3CA和PPP3CB均减少（图5B）。在鼠脑中敲除PPP3CA或PPP3R1后，SIK3 T469和S551的磷酸化均增加，但T221无变化（图5A-D）。

敲除PPP3CA或PPP3R1后，ERK1/2 T202/Y204、MEK1 S217、JNK T183/Y185、AKT1 S473、AMPKα T172、CaMK2a/b T286、PDK1 S241 或 GSK3β S9 等位点的磷酸化均无明显变化（图5A-B）。

为进一步研究CaN的底物特异性，我们进行了磷酸化蛋白质组学分析。将 PPP3R1^KD与WT^Ctrl或eGFP^Ctrl对照组进行比较后发现，在8117和8121个已确定的磷酸化位点中，分别只有148个（1.82%）和221个（2.72%）位点的磷酸化水平出现了显著上调（图S11A-B）。在两组比较中，96个相同位点的磷酸化水平上调，包括MP2K2和IP3KA等激酶、CAC1C和SCN2A等离子通道（图S11C）。在合计81个蛋白质中发现了这些位点，其中 69 个至少含有两个典型CaN底物结合基序之一（图S11D）。

PPP3CA参与小鼠睡眠调控

为了研究内源性PPP3CA在小鼠中的生理作用，我们分析了PPP3CA^KD小鼠的睡眠表型，并将其与对照组（WT^Ctrl和eGFP^Ctrl）的表型进行了比较。在EEG或EMG的一般模式上，三种基因型无显著差异(图6L和 S13E)。

图6

在24小时内，PPP3CA^KD小鼠的总睡眠时长比eGFP^Ctrl或WT^Ctrl对照小鼠减少了大约3小时(图 6A和S12 A-D, 表 S5和 S6)。在PPP3CA^KD小鼠中，NREMS时长降低约204分钟(图6B-C，表S5和S6)。在亮期，NREMS次数无变化（图6D），但NREMS单次持续时间减少(图 6E)。在暗期，NREMS次数减少（图6D），而单次持续时间无变化（图6E）。

在亮期，REMS的时长变化较小且方向相反，与对照组相比，PPP3CA^KD小鼠的REMS增加约13分钟 (图S12E-F)。在亮期，REMS时长增加主要是因REMS次数增加（图S12G），而REMS单次持续时间无变化 (图S12H). 在暗期，REMS减少约6分钟（图S12E-F），主要因REMS次数（图S12G）和单次持续时间（图S12H）减少。

EEG功率谱分析仅在ZT15和ZT23时段显示NREMS δ波功率密度减少，而其他时间点未变(图6F-H)。REMS和NREMS转换几率增加，但Wake和NREMS转换几率减少 (图6I-J和S12K-L)。

经过6小时的睡眠剥夺后，WTC^trl和eGFP^Ctrl小鼠的NREMS（图6K）或总睡眠/觉醒时间（图S13A）逐渐恢复，而PPP3CA^KD小鼠在睡眠剥夺后的相应恢复显著减少。 SD后REMS的恢复在WT^Ctrl、eGFP^Ctrl和PPP3CA^KD小鼠之间没有显著差异 (图 S13B)。

PPP3R1参与小鼠睡眠调控

小鼠全脑的免疫印迹分析显示，PPP3R1^KD小鼠的PPP3R1蛋白表达显著降低 (图5B)。代表性的EEG和EMG见图7K，代表性的睡眠图见图S15J。

图7

在每个ZT时段，除了光暗转换前后的ZT 5，PPP3R1^KD小鼠的睡眠时长显著减少，24小时内减少至少5小时(图7A 和S14A-D, 表S7和S8)。相比eGFP^Ctrl或WT^Ctrl小鼠，PPP3R1^KD小鼠的平均NREMS时长在整个24小时内减少344分钟（图7B-C和表S8），这因NREMS单次持续时间减少（图7D），而次数未变 (图7E)。. 

在亮期，PPP3R1^KD小鼠与eGFP^Ctrl或WT^Ctrl小鼠的REMS无显著差异，而在暗期，PPP3R1^KD小鼠的REMS时长平均减少10.4分钟，主要因REMS次数减少(图S14E-H，表S7和S8)。

不同睡眠和清醒状态之间的转换几率（图7J）来自图7I和图S14K及S14L的数据。从Wake到NREMS和从NREMS到NREMS的转换几率降低，而从NREMS到Wake和从Wake到Wake的转换几率增加。

PPP3R1^KD小鼠在ZT0到ZT6的6小时睡眠剥夺后缺乏NREMS和REMS的恢复。从第8到第24小时的每个时间点均有显著差异(图7K和图S15A-B)。  我们注意到，PPP3R1^KD小鼠在ZT0到ZT6期间的睡眠时间显著少于ZT6到ZT12期间。为确保在SD期间失去足够的睡眠时长，我们在ZT6到ZT12重复了睡眠剥夺（SD）。与eGFP^Ctrl和WT^Ctrl小鼠相比，PPP3R1^KD小鼠的NREMS在SD后的回复时长显著减少（图S15F），而REMS未受影响 (图S15G)。

PPP3R1^KD小鼠的NREMS δ波功率降低，而α波功率增加 (图7F-H和S14I-J)。SD后，PPP3R1^KD小鼠的NREMS δ波功率恢复也少于对照小鼠 (图 S15C-D 和H)。PPP3R1^KD小鼠在ZT6-12期间SD后的NREMS δ功率密度反弹与两个对照组无显著差异 (图S15I)。

在确认CaN在小鼠睡眠调节中的作用后，我们进一步分析了CRISPR-Cas9介导的PPP3CA或PPP3R1 KD在小鼠大脑中的效率。免疫组化分析显示，与eGFP^Ctrl和WT^Ctrl小鼠相比，PPP3CA^KD或PPP3R1^KD小鼠的大脑皮层、丘脑、中脑、后脑和下丘脑中PPP3CA或PPP3R1蛋白水平显著降低 (图S16A-H)。为确定丢失CaN的细胞类型，我们进行了病毒标记mScarlet和神经元标记Neurotrace（图S17A-C）或星形胶质细胞标记胶质纤维酸性蛋白（GFAP）（图S17E-G）的免疫组化共染色。结果显示，病毒主要侵染了神经元 (图S17D和 S17H)，这表明CaN的丢失主要发生在敲除组的神经元中。

CaN LOF的表型比SIK3 LOF的表型显著得多。为研究SIK3 T469磷酸化是否位于CaN调控睡眠的下游，我们在Sik3^T469A/+背景下敲低了PPP3R1。

在Rosa26^Cas9/+::Sik3^+/+背景下，与注射了AAV-sgRNA^EGFP的小鼠相比，PPP3R1^KD导致NREMS时长、NREMS δ波功率密度和SD后的NREMS反弹（从ZT6-12）减少(图8A-H和S19B)。在Rosa26^Cas9/+::Sik3^T469A/+背景，NREMS的α波功率（图8G）、SD前后的NREMS δ波功率密度（图8H和图8I），以及REMS到NREMS的转换几率（图8J-K）有部分回复，而总睡眠时长和NREMS时长以及NREMS在SD后的反弹表型未回复 (图8A-E和S19B)。在Rosa26^Cas9/+::Sik3^+/+小鼠及其同窝Rosa26^Cas9/+::Sik3^T469A/+小鼠中，PPP3R1^KD不影响夜间REMS时长（图S18E-H），而在后者中，它增加了ZT6-12期间SD导致的REMS反弹 (图S19C)。

图8

一些PPP3R1^KD表型，如NREMS时长和NREMS在SD后的反弹，无法通过SIK3-T469A来回复，这可以通过PPP3R1能够去磷酸化鼠脑中的其他蛋白质（如我们图S11中的那些）来解释。换句话说，SIK3不可能是CaN介导其调节睡眠的唯一底物。

讨论

我们展示了生化纯化和化学生物学在发现睡眠相关重要分子的有效性。遗传学方法虽然强大，但需要用完整的动物来发现与睡眠相关的基因。果蝇睡眠和小鼠睡眠相关的基因存在物种差异，使得有必要使用小鼠来寻找调节哺乳动物睡眠的分子，但显著减缓了寻找影响哺乳动物睡眠的分子（内源性或外源性）的过程。生物化学可以在体外应用于分子、而非动物，因此可以在更小的反应体积中使用更多的分子。用生物化学发现的分子可以在体内测试其对睡眠的影响，正如我们在这里所做的。体外分子方法可能会促进睡眠研究，并改变寻找影响睡眠药物的方法。

我们的研究结果表明，CaN在调节睡眠--特别是NREMS--起重要作用，部分通过SIK3 T469去磷酸化，并揭示了一个涉及蛋白激酶和磷酸酶的信号通路在睡眠调节中的作用。我们的研究还揭示了CaN在体外和体内去磷酸化SIK3中特定S和T残基的特异性，显示磷酸酶有之前未观察到的特异性。因SIK3敲除小鼠¹⁹和SIK3 T469A点突变小鼠的睡眠表型都比CaN突变的弱得多，很明显SIK3及其T469和S551位点并非CaN的唯一下游靶点。

我们的Sik3^S551A/+小鼠睡眠表型远弱于之前其他人报道的结果¹⁴。事实上，野生型对照小鼠的基础NREMS时长也有所不同（我们研究中的约520分钟/天 vs 他们研究中的约650分钟/天）。这种差异可能因不同的遗传背景、饲养和测量条件所造成。 Sik3^T469E/+ 小鼠无睡眠表型，这与之前的生化结果一致，即 SIK3 T469E未能模拟持续磷酸化，且无法结合14-3-3蛋白⁶⁶。Sik3^S551D/+ 小鼠同样表现出与Sik3^S551A/+ 小鼠相似、而非相反的睡眠表型¹⁴。

虽然我们和其他研究人员已经研究了许多与果蝇睡眠相关的基因，但尚未发现它们的鼠类同源基因对小鼠睡眠都是必需的。目前尚无法预测在任何一种物种中发现的基因是否会在另一种物种中调节睡眠。因此，了解到CaN参与调节果蝇睡眠是令人欣慰的^83,84。SIK3在果蝇和小鼠中都参与调节睡眠的事实^14,18,19进一步支持了SIK3及其调节因子如LKB1和CaN在睡眠控制中的重要性。

CaN在哺乳动物睡眠调节中的作用

CaN的三个催化亚基在体外都能在生化上去磷酸化SIK3的T469和S551，但不能去磷酸化T221。在体内，我们已证明PPP3CA和PPP3CB在HEK293T细胞中去磷酸化T469和S551，以及PPP3CA在小鼠脑中的作用。我们已发现PPP3CA和PPP3R1在调节小鼠睡眠方面具有生理作用。 PPP3CB的作用仍需进一步研究。

Sik3^T469A/+的部分回复效果表明CaN调节睡眠可能涉及其他下游靶点。基于CaN的两个经典底物结合基序（LxVP和PxIxIT），Wigington及同事发现了691种独特的蛋白质，这些蛋白质至少包含这两个基序中的一个，可能作为CaN的潜在底物⁸²。我们的磷酸化蛋白质组学数据还揭示了部分在CaN KD组中磷酸化状态发生变化的候选蛋白，其中离子通道蛋白（Cacna1c、Cacna1h、Kcnb1、Scn2a等）尤为引人注目。

有研究表明，SIK3在谷氨酸能神经元调节小鼠的睡眠⁸⁵。谷氨酸能神经元在小鼠的许多脑区中广泛分布⁸⁶。很可能CaN也在谷氨酸能神经元中发挥作用，从而影响睡眠。CaN 还在位于丘脑皮层和丘脑网状核的 GABA 能神经元中发挥作用⁸⁷。我们的免疫组化数据表明CaN的丢失主要发生在神经元中，但需在不同的脑区和神经元类型中系统性地操纵CaN，以找到CaN调节睡眠的主要区域。

我们发现CaN在睡眠中具有特定表型，这将激发对其他磷酸酶在睡眠中作用的进一步研究。例如，虽然CaN在NREMS中似乎比在REMS中更重要，但是否有其他磷酸酶在调节REMS方面更为重要？我们发现CaN在去磷酸化特定位点和调节睡眠特定成分方面的特异性，既带来了问题，也激发我们进一步理解哺乳动物睡眠分子机理的热情。应进一步研究磷酸酶在重要生理过程中的具体作用。

钙离子对睡眠的调节

不同脑区的Ca²⁺成像研究已显示出在不同的睡眠/觉醒状态下细胞内和细胞外的Ca²⁺浓度是不同的^88-91。利用药理学和遗传学方法操纵影响Ca²⁺浓度的离子通道可以改变小鼠的睡眠模式^38,92-95。

我们对CaN作用的发现提供了Ca²⁺的一个可能下游组分，但仍需进一步研究Ca²⁺、激酶、磷酸酶、离子通道和转录因子如何相互作用从而调节睡眠。

蛋白激酶与睡眠

第一个发现的与睡眠调节相关的激酶是PKA。一种能够抑制磷酸二酯酶并增加环腺苷酸（cAMP）的抗抑郁药可以增加大鼠的清醒度²⁶。在小鼠中，过表达一种显性负性的PKA突变体促进了REMS和NREM的碎片化，同时减少了SD后的睡眠反弹³⁰。对小鼠神经元中的ERK1或ERK2进行药理抑制或基因敲除显著减少睡眠时长³³。当发现AMPK的抑制剂会减少小鼠的睡眠，而其激活剂会增加小鼠的睡眠时，AMPK被认为与睡眠调节有关³⁴。在特定脑区通过药理学抑制CaMKII可以增加睡眠³⁷，而胚胎敲低CaMK2a和CaMK2b基因导致睡眠时间分别减少了大约50分钟和120分钟³⁸。通过小鼠的前向遗传筛选发现，SIK3的GOF突变可以调节睡眠¹⁴。在Sik3^sleepy突变体中删除了一个小片段，导致缺少PKA靶位点S551²⁰。SIK1和SIK2中的SIK3 S551同源位点的突变导致了GOF表型²¹，这就不清楚哪种SIK激酶或磷酸化位点在生理上为睡眠调节所必需。我们对每个SIK基因缺失的小鼠的研究表明，只有SIK3、而不是SIK1或SIK2，为小鼠睡眠所必需¹⁹。Lkb1是一个肿瘤抑制基因，其产物被发现能够磷酸化AMPK的α亚基的T172位点^97-103，以及包括SIK3在内的AMPK相关激酶（ARKs）的相应位点¹⁰⁴。我们最近的体内功能研究表明，LKB1为果蝇和小鼠睡眠所必需¹⁸。除了LKB1，我们体外生化研究还发现了二十多种STE20亚家族的激酶可以作为ARKs上游^105,106，这使LKB1在睡眠调节中位于SIK3上游的简单情景变得不确定。我们仍在研究这些通过生化方法鉴定的任何STE20激酶是否参与睡眠调节。

参考文献

    1. Keene, A.C., and Duboue, E.R. (2018). The origins and evolution of sleep. J Exp Biol 221，JEB159533.

附图1

附图2

附图3

附图4

附图5

附图6

附图7

附图8

附图9

附图10

附图11

附图12

附图13

附图14

附图15

附图16

附图17

附图18

附图19

附表1

附表2

附表3

附表4

附表5

附表6

附表7

附表8

FIGURE LEGENDS

Figure 1. Sleep Phenotype of T469A Males.

(A-B)Profiles showing sleep time each hr in minutes/hr (min/hr) (A) or profiles of NREMS (B). The X axis shows zeitgeber time (ZT) with the white box indicating light phase (or daytime) and black box dark phase (or nighttime). The black line shows data from SIK3^+/+ mice (n = 11), the blue line data from SIK3^T469A/+ mice (n = 10). (C-E) NREMS duration (C), NREMS episode number per 24 hr or 12 hr in light/dark phase (D), and NREMS episode duration (E). (F) EEG power spectrum during NREMS. X-axis indicates frequency distribution of EEG power. (G) Normalized EEG power of δ, θ, α, σ and β waves during NREMS. (H) Diurnal NREMS delta power density. (I) Transition probabilities. W: Wake, NR: NREMS, R: REMS. (J) A diagrammatic illustration of transition probabilities of different sleep and wake states, summarized from data in Fig. 1H and Fig. S1K and S1L. (K) Recovery of NREMS after 6 hrs of sleep deprivation (SD, ZT0-6). (L) 1 hr representative EEG and EMG traces at different vigilance states (Wake, NREMS, REMS). ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± standard error of the mean (mean ± SEM). Two-way ANOVA (A, B, I, J); One-way ANOVA (C); Kruskal-Wallis test (D, E); Two-way ANOVA with Sidak`s post-hoc test (G); Mixed-effects model (H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (F, K). 

Figure 2. Biochemical Purification of SIK3 Phosphatases from HEK Cells.

(A)A schematic illustration of phosphatase purification with HEK cell lysates passing through Q HP, Blue HP, SP HP, heparin HP, HAP HP and Superdex 200 columns before silver staining and MS analysis. (B)SIK3 T469 and S551 dephosphorylation of fractions separated by Q-HP anionic chromatography column. 500 mg (at a concentration of 10 mg/ml) HEK cell lysates flowed through 0.45 mm filter was fractionated on a Q-HP anionic chromatography column, eluted with a linear gradient of NaCl (0-600 mM) into 20 column volumes (CVs). We used one CV per fraction as the standard fraction in our purification. A small aliquot was analyzed for of phosphatase activities with SIK3. (C)SIK3 T469 and S551 dephosphorylation of fractions separated by Blue-HP column. Fractions 10 to 13 from (B) were combined, dialyzed with buffer A and loaded onto a Blue-HP column. (D)Dephosphorylation activity of fractions separated by SP-HP column. The FL fraction from (C) was further fractionated on a SP-HP column. (E) Dephosphorylation activity of fractions separated by heparin column. The FL fraction from (D) was further fractionated on a heparin column. (F) Dephosphorylation activity of fractions separated by HAP-HP column. The FL fraction from (E) was further fractionated on a HAP-HP column. The rest of the fractionation was similar to (B) except that the final wash was with 5 CVs of 500 mM K₂PO₄, giving rise to samples 21 to 25. (G) Active fractions from the HAP-HP were fractionated on a Superdex 200 column 10/300 GL, eluted into 20 fractions. 1 ml from each fraction was collected and labeled as samples 1 to 20. Protein contents were monitored with UV at 280 nm. (H) Silver-stain result and corresponding dephosphorylation activity of Fractions 10 to 18 separated by Superdex 200 column on SIK3 T469 and S551. (I) Phosphatases detected by MS.

Figure 3. Proteins Interacting with SIK3 Identified by Photo-Crosslinking.

(A)A schematic diagram of photo crosslinking method. Brain slices were prepared from SIK3-3xHA mice and photocatalytic crosslinking was carried out in darkness with eosin Y (50 mM) and 1, 6 dihexamine (1 mM). The anti-HA antibody was used to pull down proteins crosslinked to SIK3. (B)Validation of photocatalytic crosslinking efficacy in mouse brain slices. 15 min green light was given 1 hr after photocatalytic reagents treatment and slices were collected and homogenized. The arrowhead indicates SIK3-interaction protein complexes. (C-D) Veen (C) and volcano (D)plot of putative SIK3 interaction proteins enriched in photo-linkage groups. The significant threshold in (D) was set at P value < 0.05 and Fold change > 4. Blue dots represented protein candidates reaching significant threshold. PPP3CA and PPP3R1 were highlighted in red. Members of 14-3-3 family and one PKA catalytic subunit PRKACB were highlighted in black. (E-F) Co-immunoprecipitation between PPP3CA and SIK3 from WT mouse brain homogenates using anti-SIK3 and anti-PPP3CA antibodies. (G-H) Co-immunoprecipitation between the HA tagged PPP3CA and FLAG tagged SIK3 in HEK293T cells using anti-FLAG and anti-HA antibodies. HEK293T cells were co-overexpressed with 1 μg HA-PPP3CA and 1 μg FLAG-SIK3 expressing plasmids for 24 hrs before cell collection and lysis.

Figure 4. Site-Specific Dephosphorylation of SIK3 by PPP3CA in vitro and in vivo. 

(A) Dephosphorylation activity of PPP3CA, PPP3CB, PPP3CC and PPP5C. Each PPases, PPP3R1 and full length SIK3 were separately immunoprecipitated from HEK cells. (B-E)PPP3CA dephosphorylation of MEK1 at S217, AKT1 at T308, PDK1 at S241, MARK1 at T215, GSK3β at S9, HDAC4 at S246 and S632, JNK1 at T183/Y185, ERK2 at T185/Y187. Proteins were separately immunoprecipitated from HEK cells. (F) Subunits and cation dependence of calcineurin dephosphorylation of SIK3 T221, T469 and S551 in vitro. Recombinant PPP3CA, PPP3R1, CaM were purified from E. coli. Full-length SIK3 was immunoprecipitated from HEK 293T cells. (G-H) Dose-dependent dephosphorylation effect of ionomycin on SIK3 T469 and S551 in HEK cells. (I-J) Time-dependent dephosphorylation effect of ionomycin on SIK3 T469 and S551 in HEK cells. (K-L) Alteration of ionomycin dephosphorylation effect on SIK3 T469 and S551 by PPP3CA/PPP3R1 co-transfection. Numbers represent plasmid gram weight. After transfection for 24 hrs, cells were treated with 10 μg/ml ionomycin for 10 min. (M) Alteration of ionomycin dephosphorylation effect on SIK3 T469 and S551 by PPP3CA-H151A/PPP3R1 co-transfection. (N-P) Alteration of ionomycin dephosphorylation effect on SIK3 T469 and S551 by PPP3CA/PPP3CB double KO or PPP3R1 KO. ns, not significant; *p < 0.05; **p <0.01; ***p <0.001; mean ± standard error of the mean (mean ± SEM). One-way ANOVA (H, J, L); Student`s t-test (O, P).

Figure 5. Serine/Threonine Phosphorylation in the Mouse Brain after PPP3CA or PPP3R1 Knockdown.

(A) PPP3CA knockdown caused serine/threonine phosphorylation alteration in the brain. Each lane shows results from one mouse. From the left are: two wt mice injected with sgRNAs targeting GFP, five Cas9 expressing mice injected with sgRNAs targeting GFP, and five Cas9 expressing mice injected with sgRNAs targeting PPP3CA. (B) PPP3R1 knockdown caused serine/threonine phosphorylation alteration in the brain. From the left are: three WT mice injected with sgRNAs targeting GFP, three Cas9 expressing mice injected with sgRNAs targeting GFP, and three Cas9 expressing mice injected with sgRNAs targeting PPP3R1. (C-D) Statistical analysis for (A) and (B). ns, not significant; *p < 0.05; **p <0.01; mean ± standard error of the mean (mean ± SEM). One-way ANOVA with Tukey's multiple comparisons test.

Figure 6. Sleep Phenotype of Mice after PPP3CA Knockdown.

(A-B)Profiles of total sleep (A) or NREMS (B) in PPP3CA^KD, eGFP^Ctrl and WT^Ctrl mice. (C)NREMS duration over 24 hrs. (D-E) NREMS episode number per 24 hr or 12 hr in light/dark phase (D) and episode duration (E). (F) EEG power spectrum during NREMS. (G) Normalized EEG power of δ, θ, α, σ and β waves during NREMS. (H) Diurnal NREMS delta power density. The jump points at ZT12-15 and ZT23 were due to lack of data because most mice in PPP3CA^KDgroup were awake at these time points. (I) NREMS related transition probabilities. (J) A diagrammatic illustration of transition probabilities of different sleep and wake states. (K) Recovery of NREMS after 6 hrs of SD (ZT0-6). (L) 1 hrrepresentative EEG and EMG traces at different vigilance states. ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± standard error of the mean (mean ± SEM). Two-way ANOVA (A, B, I, J); One-way ANOVA (C); Kruskal-Wallis test (D, E); Two-way ANOVA with Sidak`s post-hoc test (G); Mixed-effects model (H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (F, K).

Figure 7.Sleep Phenotype of Mice after PPP3R1 Knockdown.

(A-B) Profiles of total sleep (A) or NREMS (B) in eGFP^Ctrl, WT^Ctrl and PPP3R1^KD mice. (C) Data and statistics of NREMS duration over 24 hrs. (D-E) NREMS episode number per 24 hr or 12 hr in light/dark phase (D) and episode duration (E). (F) NREMS EEG power spectrum analysis. (G) Normalized EEG power of δ, θ, α, σ and β waves during NREMS. (H) NREMS delta power density over 24 hrs. (I) NREMS related transition probabilities. (J) Transition probabilities of different sleep and wake states. (K) Recovery of NREMS after 6 hrs of SD (ZT0-6). (L) 1 hrrepresentative EEG and EMG traces at different vigilance states. ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± standard error of the mean (mean ± SEM). Two-way ANOVA (A, B, I, J); One-way ANOVA (C); Kruskal-Wallis test (D, E); Two-way ANOVA with Sidak`s post-hoc test (G); Mixed-effects model (H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (F, K).

Figure 8. Partial Rescue of Sleep Phenotype by SIK3^T469A/+ Background after PPP3R1 Knockdown in Mice.

(A-B) Profiles of total sleep (A) or NREMS (B) in eGFP^Ctrl: Rosa26^Cas9/+; SIK3^+/+ mice injected with AAV-eGFP-sgRNA virus, PPP3R1^KD: Rosa26^Cas9/+; SIK3^+/+mice injected with AAV-ppp3r1-sgRNA virus and SIK3^T469A/+; PPP3R1^KD: Rosa26^Cas9/+; SIK3^T469A/+mice injected with AAV-ppp3r1-sgRNA virus. (C) NREMS duration over 24 hrs.(D-E) NREMS episode duration (D) and episode number per 24 hr or 12 hr in light/dark phase (E). (F) NREMS EEG power spectrum analysis. (G) Normalized EEG power of δ, θ, α, σ and β waves during NREMS. (H) NREMS delta power density over 24 hrs. (I) NREMS delta power density during the 24 hrs recovery after SD (ZT6-12). (J) REMS related transition probabilities. (K) Transition probabilities of different sleep and wake states. (L) 1hr representative EEG and EMG traces at different vigilance states. ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± standard error of the mean (mean ± SEM). Two-way ANOVA (A, B, J, K); One-way ANOVA (C); Kruskal-Wallis test (D, E); Two-way ANOVA with Sidak`s post-hoc test (G); Mixed-effects model (H, I); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (F).

Figure S1. Additional Sleep Phenotype in Male T469A Mutant Mice.

(A-D) Profiles of wake over 24 hrs (A), total wake duration over 24 hrs (B), wake episode number per 24 hr or 12 hr in light/dark phase (C), wake episode duration (D) of Sik3^T469A/+ (blue, n=10) and Sik3^+/+ mice (black, n = 11).

(E-H) Profiles of REMS over 24 hrs (E), total REM duration over 24 hrs (F), REM episode number per 24 hr or 12 hr in light/dark phase (G), REM episode duration (H).

(I-J) EEG power spectrum during Wake (I) or REMS (J).

(K-L) Self- (K) and REMS related (L) transition probabilities between different sleep and wake states. W: Wake, NR: NREMS, R: REMS.

ns, not significant; *p < 0.05; mean ± SEM. Two-way ANOVA (A, E, K, L); One-way ANOVA (B, F); Kruskal-Wallis test (C, D); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (I, J).

Figure S2.Additional Sleep Phenotype in Male T469A Mutant Mice.

(A-B)Recovery of Wake (A) or REMS (B) after 6 hrs of SD. ns, not significant; mean ± SEM (Two-way ANOVA with Tukey's multiple comparisons test). 

(C) NREMS delta power density during the 24 hrs recovery after SD. ns, not significant; *p< 0.05; **p <0.01; ***p<0.001; ****p <0.0001; mean ± SEM (Mixed-effects model). 

(D) Changes of NREMS delta power density after SD. ns, not significant; mean ± SEM (Two-way repeated measurement ANOVA with Tukey's multiple comparisons test). 

(E) Representative hypnograms of littermates.

Figure S3.Sleep Phenotype in Female T469A Mutant Mice.

(A-D) Profiles of NREMS over 24 hrs (A), total NREMS duration over 24 hrs (B), NREMS episode number per 24 hr or 12 hr in light/dark phase (C), NREMS episode duration (D).

(E-H) Profiles of REMS over 24 hrs (E), total REMS duration over 24 hrs (F), REMS episode number per 24 hr or 12 hr in light/dark phase (G), REMS episode duration (H).

(I-L) Profiles of wake over 24 hrs (I), total wake duration over 24 hrs (J), wake episode number per 24 hr or 12 hr in light/dark phase (K), wake episode duration (L).

(M-P) Probabilities of transition between different sleep and wake states.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, E, I, M-P); One-way ANOVA (B, F, J); Kruskal-Wallis test (C, D, G, H, K, L). 

Figure S4.Additional Sleep Phenotype in Female T469A Mutant Mice.

(A-C) EEG power spectrum during NREMS (A), REMS (B) or wake (C).

(D) Normalized EEG power of δ, θ, α, σ and β waves during NREMS.

(E) Diurnal NREMS delta power density.

(F-H)Recovery of NREMS (F), REMS (G) and Wake (H) after 6 hrs of SD (ZT0-6).

(I) NREMS delta power density during the 24 hrs recovery time.

(J) Changes of NREMS delta power density after 6 hrs of SD.

(K) 1 hr representative EEG and EMG traces of littermates at each vigilance state.

(L) Representative hypnograms of littermates.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (F, G, H); Mixed-effects model (E, I); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (A, B, C, J). Two-way ANOVA with Sidak`s post-hoc test (D). 

Figure S5.Sleep Phenotype in Male S551A Mutant Mice.

(A-D) Profiles of NREMS over 24 hrs (A), total NREMS duration over 24 hrs (B), NREMS episode number per 24 hr or 12 hr in light/dark phase (C), NREMS episode duration (D). 

(E-H) Profiles of wake over 24 hrs (E), total wake duration over 24 hrs (F), wake episode number per 24 hr or 12 hr in light/dark phase (G), wake episode duration (H).

(I-L) Profiles of REMS over 24 hrs (I), total REMS duration over 24 hrs (J), REMS episode number per 24 hr or 12 hr in light/dark phase (K), REMS episode duration (L).

(M-P) Probabilities of transition between different sleep and wake states.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, E, I, M-P); One-way ANOVA (B, F, J); Kruskal-Wallis test (C, D, G, H, K, L).

Figure S6.Additional Sleep Phenotype in Male S551A Mutant Mice.

(A-C) EEG power spectrum during NREMS (A), REMS (B) or Wake (C). 

(D) Normalized EEG power of δ, θ, α, σ and β waves during NREMS.

(E) Diurnal NREMS delta power density.

(F-H)Recovery of NREMS (F), REMS (G) and Wake (H) after 6 hrs of SD (ZT0-6).

(I) NREMS delta power density during the 24 hrs recovery time.

(J) Changes of NREMS delta power density after 6 hrs of SD.

(K) 1 hr representative EEG and EMG traces of littermates at each vigilance state.

(L) Representative hypnograms of littermates.

(M-N) Western-blot showing phosphorylation change in SIK3 T221, T469 and S551 sites in SIK3^T469A/+ (L) and SIK3^S551A/+ (M) mouse brain.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (F, G, H); Mixed-effects model (E, I); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (A, B, C, J). Two-way ANOVA with Sidak`s post-hoc test (D). 

Figure S7.Sleep Phenotype in male T469E Mutant Mice.

(A-D) Profiles of NREMS over 24 hrs (A), total NREMS duration over 24 hrs (B), NREMS episode number per 24 hr or 12 hr in light/dark phase (C), NREMS episode duration (D).

(E-H) Profiles of wake over 24 hrs (E), total wake duration over 24 hrs (F), wake episode number per 24 hr or 12 hr in light/dark phase (G), wake episode duration (H).

(I-L) Profiles of REMS over 24 hrs (I), total REMS duration over 24 hrs (J), REMS episode number per 24 hr or 12 hr in light/dark phase (K), REMS episode duration (L).

(M-P) Probabilities of transition between different sleep and wake states.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, E, I, M-P); One-way ANOVA (B, F, J); Kruskal-Wallis test (C, D, G, H, K, L).

Figure S8.Additional Sleep Phenotype in male T469E Mutant Mice.

(A-C) EEG power spectrum during NREMS (A), REMS (B) or Wake (C).

(D) Normalized EEG power of δ, θ, α, σ and β waves during NREMS.

(E) Diurnal NREMS delta power density.

(F-H)Recovery of NREMS (F), REMS (G) and Wake (H) after 6 hrs of SD (ZT0-6).

(I) NREMS delta power density during the 24 hrs recovery time.

(J) Changes of NREMS delta power density after 6 hrs of SD.

(K) 1 hr representative EEG and EMG traces of littermates at each vigilance state.

(L) Representative hypnograms of littermates.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (F, G, H); Mixed-effects model (E, I); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (A, B, C, J). Two-way ANOVA with Sidak`s post-hoc test (D).

Figure S9. Verification of AntibodySpecificities against SIK3 pT469/pS551, purification table and Dephosphorylation Activity of PPP5C.

(A)Specificity test of antibodies targeting SIK3 pT469 and pS551. Recombinant SIK3(1-558) was incubated with recombinant PKA^T197E in the presence or absence of ATP. SIK3 T469 and S551 phosphorylation were then detected using indicated antibodies.

(B) Purification table showing the enzymatic activity after each step.

(C) Verification of PPP5C dephosphorylation activity. PPP5C purified from HEK293T cells was incubated with either Tau IPed from HEK293T. Phosphorylation level of indicated sites were examined.

Figure S10.A Schematic Diagram of Virally Mediated Gene Knockdown in Mice.

Host mice were either WT or had Cas9 inserted at its Rosa26 site (Rosa26^Cas9/+). Each host mouse was injected with an AAV virus two weeks before an EEG recorder was placed on its head.

Figure S11. Phospho-proteomic Analysis of Brain Samples from PPP3R1 Knockdown Mice.

(A-B) Volcano plots showing changes of phosphopeptides in PPP3R1^KD/WT^Ctrl (A) and PPP3R1^KD/eGFP^Ctrl (B) groups. The significant threshold was set as p value < 0.05 and fold change > 2 or < -2. Numbers in red color represented the amounts of phosphopeptides upregulated and numbers in brackets meant the amounts of total phosphopeptide identified.

(C) Venn diagram of significantly upregulated phosphopeptides among two groups.

(D) Venn diagram of calcineurin binding motif distribution among 81 proteins to which common upregulated phosphopeptides among two groups belonged.

Figure S12.Additional Sleep Phenotype in Male PPP3CA Knockdown Mice.

(A-D)Profiles of wake over 24 hrs (A), total wake duration over 24 hrs (B), wake episode number per 24 hr or 12 hr in light/dark phase (C), wake episode duration (D).

(E-H)Profiles of REMS over 24 hrs (E), total REMS duration over 24 hrs (F), REMS episode number per 24 hr or 12 hr in light/dark phase (G), REMS episode duration (H).

(I-J)EEG power spectrum during Wake and REMS.

(K-L) Self- (K) and REMS related (L) transition probabilities between different sleep and wake states.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, E); One-way ANOVA (B, F); Kruskal-Wallis test (C, D, G, H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (I-L). 

Figure S13.Additional Sleep Phenotype in Male PPP3CA Knockdown Mice.

(A-B) Recovery of wake (A) and REMS (B) after 6 hrs of SD.

(C) NREMS delta power density during the 24 hrs recovery time.

(D) Changes of NREMS delta power density after 6 hrs of SD.

(E) Representative hypnograms of littermates.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, B); Mixed-effects model (C); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (D).

Figure S14.Additional Sleep Phenotype in PPP3R1 Knockdown Male Mice.

(A-D) Profiles of wake over 24 hrs (A), total wake duration over 24 hrs (B), wake episode number per 24 hr or 12 hr in light/dark phase (C), wake episode duration (D).

(E-H) Profiles of REMS over 24 hrs (E), total REMS duration over 24 hrs (F), REMS episode number per 24 hr or 12 hr in light/dark phase (G), REMS episode duration (H).

(I-J) EEG power spectrum analysis of Wake (I) and REMS (J).

(K-L) Self- (K) and REMS related (L) transition probabilities of different sleep and wake states.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, E); One-way ANOVA (B, F); Kruskal-Wallis test (C, D, G, H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (I-L).

Figure S15.Additional Sleep Phenotype in PPP3R1 Knockdown Male Mice.

(A-B) Recovery of Wake (A) and REMS (B) after 6 hrs of SD.

(C) NREMS delta power density during the 24 hrs recovery period after 6 hrs of SD.

(D) Changes of NREMS delta power density after 6 hrs of SD.

(E-G) Recovery of wake (E), NREMS (F) and REMS (G) after 6 hrs of SD from ZT6-12.

(H) NREMS delta power density during the 24 hrs recovery period after 6 hrs of SD from ZT6-12.

(I) Changes of NREMS delta power density after 6 hrs of SD from ZT6-12.

(J) Representative hypnograms of littermates.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, B, E-G); Mixed-effects model (C, H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (D, I).

Figure S16. PPP3CA and PPP3R1 knockdown in the mouse brain. (A-C) Immunohistochemical staining of PPP3CA in the brain slices from WT^Ctrl (A), eGFP^Ctrl (B) and PPP3CA^KD (C) groups. Scalebars, 2000 μm. N=3. (D-F) Immunohistochemical staining of PPP3R1 in the brain slices from WT^Ctrl (D), eGFP^Ctrl (E) and PPP3R1^KD (F) groups. Scalebars, 2000 μm. N=2-3.  (G-H) PPP3CA and PPP3R1 expression levels after semi-quantification by immunofluorescence.

Figure S17. Cell types of AAV infection within mouse brain. (A-C) Co-staining of mScarlet and Neurotrace in the brain slices from eGFPCtrl (A) and PPP3CAKD (B) and PPP3R1KD (C) groups. Scalebars, 2000 μm. N=3. The inset graphs below represent the overlap between mScarlet and Neurotrace signals. Scalebars, 100 μm. N=3. (D) The percentage of neurons in mScarlet positive cells. (E-G) Co-staining of mScarlet and GFAP in the brain slices from eGFPCtrl (E) and PPP3CAKD (F) and PPP3R1KD (G) groups. Scalebars, 2000 μm. N=3.The inset graphs on below represent the overlap between mScarlet and GFAP signals. Scalebars, 100 μm. N=3. (H) The percentage of GFAP in mScarlet positive cells.

Figure S18.Additional Rescue Results of Sleep Phenotype after Knockdown of PPP3R1 in SIK3^T469A/+Mice.

(A-D) Profiles of wake over 24 hrs (A), total wake duration over 24 hrs (B), wake episode duration (C), wake episode number per 24 hr or 12 hr in light/dark phase (D).

(E-H) Profiles of REMS over 24 hrs (E), total REMS duration over 24 hrs (F), REMS episode duration (G), REMS episode number per 24 hr or 12 hr in light/dark phase (H).

(I-J) EEG power spectrum analysis of Wake (I) and REMS (J).

(K-L) Self- (K) and NREMS related (L) transition probabilities of different sleep and wake states.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A, E); One-way ANOVA (B, F); Kruskal-Wallis test (C, D, G, H); Two-way repeated measurement ANOVA (Two-way RM ANOVA) (I-L).

Figure S19.Additional Rescue Results of Sleep Phenotype after Knockdown of PPP3R1 in SIK3^T469A/+ Mice.

(A-C) Recovery of Wake (A), NREMS (B) and REMS (C) after 6 hrs of SD from ZT6-12.

(D) Changes of NREMS delta power density after 6 hrs of SD from ZT6-12.

(E) Representative hypnograms of littermates.

ns, not significant; *p < 0.05; **p <0.01; ***p <0.001 and ****p <0.0001; mean ± SEM. Two-way ANOVA (A-C);Two-way repeated measurement ANOVA (Two-way RM ANOVA) (D).

SUPPLEMENTAL INFORMATION

Supplemental information includes 19 figures and 8 tables.

Key Resources Table

Methods

Mouse stocks

WT C57 BL/6J mice (8 to 10 weeks old) were purchased from Beijing Vital River Laboratories Technology Co., Ltd. or Laboratory animal resource center in Chinese Institute for Brain Research. Rosa26-Cas9 knock-in (Rosa26^Cas9/+, RRID: IMSR_JAX:024858) mice were obtained from Jackson laboratory (Platt et al., 2014). SIK3-3xHA mice were constructed in BIOCYTOGEN by infusing a 3xHA-T2A-iCre cassette into the C terminus right before the SIK3 stop codon. SIK3 point mutant mice were constructed with CRISPR-Cas9 mediated homologous recombination. For SIK3^T469A/+, the gRNA sequence was 5’- TTTGTCAATGAGGAGGCACA-3’ and a single strand homologous arm was designed to introduce nucleotide mutation from ACG to GCG (T469A) as well as a restriction enzyme site BstUI for future genotyping, sequence of which was 5’-CCTTCTCCAGAAGCCTTGGTTCGCTATTTGTCAATGAGGAGGCACGCGGTGGGAGTGGCTGACCCACGGTAAGTACCTGGTCAGCATCCT-3’. For SIK3^T469E/+, the gRNA sequence was 5’-TTTGTCAATGAGGAGGCACA-3’, the single strand homologous arm with BstUI was 5’-CCTTCTCCAGAAGCCTTGGTTCGCTATTTGTCAATGAGGAGGCATGAAGTGGGAGTGGCTGACCCACGGTAAGTACCTGGTCAGCATCCT-3’. For SIK3^S551A/+, the gRNA sequence was 5’-GGCCGGAGAGCCTCAGATGG-3’, the single strand homologous arm with BstUI was 5’-ACACTACAGCTACTGAACGGAATGGGGCCCCTTGGCCGGAGAGCCGCGGACGGAGGCGCCAACATCCAACTGCATGCCCAGCAGCTGCTCAAG-3’. A mixture of Cas9-expressing mRNA, single strand homologous arm and sgRNA was injected into fertilized eggs through electroporation and the eggs were then transplanted into the womb of foster mothers. F0 and F1 mice were genotyped through PCR and BstUI digestion to make sure the presence of recombination. Mutant lines were back-crossed to C57BL/6J for at least 5 generations to exclude possible off-targeting.

Mouse housing

All experimental procedures were performed in accordance with the guidelines and were approved by the Animal Care and Use Committee of Chinese Institute for Brain Research, Beijing. Mutant mice and wt littermates were maintained on a C57 BL/6J background. Mice were housed under a 12 hr:12 hr light/dark cycle and controlled temperature and humidity conditions. Food and water were delivered ad libitum. Mice used in all experiments were 10-14 weeks old.

Cell Culture and cDNA Transfection

HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco, C11995500BT) medium containing 10% fetal bovine serum (Transgen, FS101-02) and 1% Penicillin/Streptomycin (Gibco, 15070-063). cDNAs were transfected into HEK293T cells with Lipofectamine 3000 reagent (Thermo Fisher, L3000015) according to the manufacturer’s instructions and harvested 24 to 28 hrs after transfection.

Drug Treatment and Protein Preparation

HEK293T cells were treated with ionomycin (MCE, SQ23377) at indicated concentrations and time durations at 37℃. Cells were then harvested and lysed with 1 ml lysis buffer (0.3% Chaps, 10 mM KCl, 1.5 mM MgCl₂, 1 mM EDTA, 1 mM EGTA, pH 7.4, 1x protease inhibitor cocktail (Roche, cOmplete™), 1x phosphatase inhibitor II (Sigma, P5726) and 1x phosphatase inhibitor III (Sigma, P0044)) before centrifugation of cell lysates at 13000 rpm for 10 mins at 4 ℃. Protein concentrations of cell lysates were determined with the bicinchoninic acid assay (Thermo Fisher, 23225) and normalized to 2 mg/ml. Samples were analyzed by immunoblotting with the indicated antibodies.

Mouse Brain Protein Preparation

Whole brains of mice were quickly dissected, rinsed withP BS and homogenized by homogenizer (Wiggens, D-500 Pro) in ice-cold lysis buffer (150 mM NaCl, 1% Triton-X-100, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-base, freshly supplemented with a protease and phosphatase inhibitor cocktail). Brain homogenates were centrifuged at 15000 rpm for 25 mins at 4℃. Supernatants were carefully transferred into a new microtube. Protein concentrations of brain lysates were determined with the bicinchoninic acid assay and normalized to 2 mg/ml. Before immunoblotting, samples were kept in liquid nitrogen, if necessary. 

Immunofluorescence and microscope

Mice were anesthetized with 3% Tribromoethanol and perfused with 1× PBS followed by chilled 4% paraformaldehyde (PFA). Then the mouse brain was dissected. After 24h post-fixation at 4℃ and gradient dehydration from 20% sucrose to 30%, 40-μm sagittal brain slices were collected sequentially using a cryostat (CM3050 S, Leica). Free-floating sections were washed with PBS and then blocked in 5% albumin bovine serum (BSA, Sigma) in PBST (0.5% Triton X-100 in PBS) at room temperature for 2 h. After blocking, the sections were incubated with primary antibodies diluted in 3% albumin bovine serum (BSA, Sigma) in PBST (0.3% Triton X-100 in PBS) at 4℃ overnight. The primary used were as follows, rabbit anti-Calcineurin antibody (1:500, Abcam), rabbit anti-CNB antibody (1:500, Abcam), Chicken anti-GFAP antibody (1:500, Abcam), rabbit anti-RFP antibody (1:1000, Rockland). Then the sections were washed for 3 times with PBS before incubated in 3% albumin bovine serum with secondary antibody for 2h. Alexa 546 goat anti-rabbit IgG (H + L) (1:500, Invitrogen), alexa 488 goat anti-rabbit IgG (H + L) (1:500, Invitrogen), alexa 488 goat anti-chicken IgG (H + L) (1:500, Invitrogen) was used as the secondary antibody. Neurotrace 505/525 (1:500) was used. After 3 times wash, the brain slices were attached to adhesive slides. Images were collected using LSM880 confocal microscope (Zeiss) with 10×/0.45NA air objective and a 20×/0.8NA air objective. Mean fluorescence intensity was calculated for semi-quantitative fluorescence analysis using the software imageJ. Neuron positive cells were also counted using imageJ.

Biochemical Purification

Lysates from HEK293T cells were prepared and filtered through 0.45 mm filters (Millipore, SLHV033R). 500 mg cell lysates at a concentration of 10 mg/ml were fractionated on a Q HP (Cytiva, 17115401) anionic chromatography column, eluted with a linear gradient of NaCl (0-600 mM) into 20 column volumes (CVs), with each fraction collected as one CV as samples 1 to 20, and the final wash with 1 M NaCl buffer A (20mM, HEPES, 10mM KCl, 1.5mM MgCl₂, 3mM DTT, pH 7.4) with 5 CVs gave rise to samples 21 to 25. Each fraction was dialyzed into buffer A, removing NaCl. 10 ml of each sample was used for analysis of activities removing phosphate from SIK3 T469 and S551. T469 and S551 of bacterially expressed recombinant SIK3 were phosphorylated by PKA in vitro before being used to test phosphatase activities of the fractions of HEK lysates. Fractions 10 to 13 contained significant activities removing phosphate from SIK3 T469 and S551.Fractions 10 to 13 from QHP were combined and dialyzed with buffer A before being loaded onto a Blue HP column (Cytiva, 17041301). It was eluted with a linear gradient of NaCl (0-600 mM) into 20 column volumes (CVs), with each fraction collected as one CV as samples 1 to 20, and the final wash with 1 M NaCl buffer A with 5 CVs gave rise to samples 21 to 25. The FL fraction from the Blue HP contained significant activities removing phosphate from SIK3 T469 and S551.The FL fraction from the Blue HP column was dialyzed with buffer A and loaded onto a SP HP column. The FL fraction from the SP-HP column (Cytiva, 17115201) contained significant activities removing phosphate from SIK3 T469 and S551.The FL fraction from the SP-HP column was dialyzed with buffer A and loaded onto a heparin HP column (Cytiva, 17040703). The FL fraction and Fraction 1 contained significant activities removing phosphate from SIK3 T469 and S551.The FL fraction from the heparin HP column was dialyzed with buffer A and loaded onto a HAP HP column (BioRad, 7510025). The rest of the fractionation was similar to the first column except that the final wash was with 5 CVs of 500 mM K₂PO₄, giving rise to samples 21 to 25. Fractions 5 to 7 contained significant activities removing phosphate from SIK3 T469 and S551.Active fractions from the HAP HP column were condensed into 0.5 ml, fractionated on a Superdex 200 molecular sieve column (Cytiva, 28990944), eluted with 200 mM NaCl into 20 ml. 1 ml from each fraction was collected and labeled as samples 1 to 20. Protein contents were monitored with UV at 280 nm.

Photocatalytic Crosslinking

Because green LED irradiation significantly initiates the crosslinking reaction which leads to covalent capture of PPIs, this strategy allows us to avoid the loss of interacting proteins caused by sample processing steps such as washing during subsequent enrichment and MS analysis. It thus reduces false-negative results caused by traditional PPI study methods such as affinity purification-mass spectrometry (AP-MS). Therefore, in our experiments, we prepared samples treated with and without green light irradiation treatment simultaneously, of which the latter served as background results of normal immunoprecipitation as no photo-crosslinking reaction occurs. By quantitatively comparing the enriched interacting proteins between the +GL and -GL groups with MS analysis, proteins significantly enriched in the light-irradiated group become primary candidates.

SIK3-3xHA mice were sacrificed and brain slices were kept in PBS. Neurobasal medium (Gibco, 21103049) supplied with B-27 (Gibco, 17504044) for photocatalytic crosslinking was prepared in darkness by adding eosin Y (Sigma-Aldrich, E4009) and 1,6-dihexamine (Sigma-Aldrich, H11696) to a working concentration of 50 μM and 1 mM, respectively. Cell culture chambers (Millipore, PICM0RR50) were placed in 6-well plates and rinsed with PBS. Brain slices were transferred to chambers carefully with a sterile dropper and each chamber finally contained four slices to ensure thorough stretch of each slice. PBS was discarded by pipetting from the outer side of chambers and 0.5 ml of neurobasal medium containing photocatalytic crosslinkers was added into each well to ensure that each slice was totally infiltrated with the probes. Samples were incubated at 37°C with 5% CO₂ for 1 hr before photo-irradiation. For photocatalytic crosslinking, the plates were placed on green LED (520 nm, 20 mW/cm²) equipment while an ice bag and a fan were used to reduce light irradiation-induced heat. After 15 mins of green light irradiation, the color of the medium was bleached, indicating effective activation of the eosin probe. The medium was discarded and after one round of PBS wash, the slices was collected into 1.5 ml Eppendorf tubes, which were placed into liquid N₂ to freeze the slices.

Enrichment of Crosslinked Proteins

To each tube with frozen brain slice samples was added 1 ml of ice-cold lysis buffer containing 1% of protease cocktail and a steel ball. Samples were lysed through ultrasonication. After centrifugation (12,000 g, 10 mins, 4°C) to discard the residue, crosslinked proteins were enriched by anti-HA magnetic beads (Pierce, 88837) according to the manufacture protocol. Importantly, the beads should be washed with lysis buffer (with 0, 0.25, and 0.5 M NaCl) for three times to diminish non-specific binding proteins. Crosslinked proteins were eluted with 2x SDS-loading buffer and heated to 95 °C for 10 mins. Eluted proteins were subjected to further Western analysis and LC-MS/MS.

In-gel digestion

Proteins enriched with HA beads were loaded on an 8% SDS-PAGE gel and run at 150 V for 30 mins. After silver staining (Sigma-Aldrich, PROTSIL2), the desired bands of protein mixture were excised and cut into 1 mm³ pieces. The gel pieces were discolored in discoloring buffer until they all turned transparent. A dehydration process was carried out by adding pure acetonitrile into gel pieces until they were totally dehydrated to appear non-transparently white. Samples were then incubated in the reduction buffer (10 mM DTT (Sigma-Aldrich, D9779), 50 mM ammonium bicarbonate (Sigma-Aldrich, A6141)) at 56 °C for 30 mins and further incubated in alkylation buffer (55 mM iodoacetamide (Sigma-Aldrich, I6125), 50 mM ammonium bicarbonate) at 37 °C for 30 mins in the dark. After washed by 50 mM ammonium bicarbonate buffer twice, gel pieces were dehydrated through the same protocol. 20 ng/μltrypsin buffer in 50 mM ammonium bicarbonate was added and samples were incubated at 4 °C for one hr. The remaining buffer was discarded and 50 mM ammonium bicarbonate buffer was added for another 16 hrs of digestion at 37 °C. The resulting peptides were extracted with extraction buffer (50% acetonitrile, 45% water and 5% formic acid (Honeywell, 09676)), before being centrifuged to dryness under vacuum.

Dimethyl labeling

The collected peptides were reconstituted in 100 mM TEAB buffer (Sigma-Aldrich, T7408). 39.688 mg/mL of NaBH₃CN (Sigma-Aldrich, I56159) followed by 4% (v/v) CH2O or CD2O were added for light and heavy dimethyl labeling, respectively, following the addition of 39.688 mg/ml NaBH₃CN. After incubation in a fume hood for 30 mins at room temperature, enough 1% (v/v) ammonia solution and formic acid were added immediately to quench the labeling reaction. The light and heavy samples were combined together, and then desalted and dried under vacuum.

LC-MS/MS analysis

Trypsin digested peptides were analyzed on a Exploris 480 Hybrid Quadrupole Orbitrap Mass Spectrometer as well as Thermo Scientific Q Exactive Orbitrap Mass Spectrometer in conjunction with an Easy-nLC II HPLC (Thermo Fisher Scientific). The mobile phases were A: 0.1% formic acid in H₂O; B: 0.1% formic acid in 80% ACN–20% H₂O. MS/MS analysis was performed under the cationic mode with a full-scan m/z range from 350 to 1,800 and a mass resolution of 70,000.

Peptide identification

For quantitative SIK3 interactome analysis, the quantification of light/heavy ratios was calculated with a precursor mass tolerance of 20 ppm. Alkylation of cysteine (+57.0215 Da) was set as the static modification, and oxidation of methionine (+15.9949 Da) and acetylation of N-terminal Lys (+42.0106 Da) was assigned as the variable modification. The isotopic modifications (28.0313 and 32.0557 Da for light and heavy labeling, respectively) were set as fixed modifications on the peptide N-terminus and lysine residues. Half-tryptic terminus and up to two missing cleavages were set within tolerance.

Co-immunoprecipitation

For HEK293T cells, plasmids expressing FLAG-tagged SIK3 and HA-tagged PPP3CA were transfected for 24 hrs. Cells were then collected and lysed with 1 ml lysis buffer (25 mM Tris-base pH 7.4, 150 mM NaCl, 1% NP40, 1mM EDTA, 5% glycerol,1x protease inhibitor cocktail, 1x phosphatase inhibitor II and 1x phosphatase inhibitor III) before centrifugation at 13000 rpm for 10 mins at 4℃. 40 μl supernatants were transferred into new microtubes as input samples and the rest was incubated either with 20 μl anti-FLAG (Millipore, M8823) or anti-HA antibody coated magnetic beads balanced by lysis buffer for 1 hr at 4 ℃. Beads were then washed with 1 ml lysis buffer three times at 4 ℃ and 40 μl PBS was added to transfer the beads into new microtubes as enriched samples. For mouse, anti-PPP3CA or anti-SIK3 antibody and corresponding IgG were pre-incubated with protein A/G beads (YEASEN, 36401ES08) for 2 hrs at 4℃ in lysis buffer. Protein samples from mouse brain were at first pre-absorbed using IgG-protein A/G beads for 30 mins at RT and 40 μl supernatants were transferred into a new microtube as input sample before anti-PPP3CA/SIK3 protein A/G beads was added and rotated overnight at 4℃. Beads were then washed with 1 ml lysis buffer three times at 4℃ and 40μl PBS was added to transfer the beads into a new tube as enriched samples.

Expression of Recombinant Proteins inE. coli

cDNAs for specific proteins were subcloned into the pET-28a vector, with appropriate tags such as MBP, GFP or FLAG. Plasmids were transfected into E. coli BL21 and incubated at 37 ℃ until the OD reached 0.6, when 0.5 mM of IPTG (Sigma, I6758) was added at 18 ℃ to induce protein expression for 16 hrs. Cells were collected and treated with Ni column binding buffer (300 mM NaCl, 20 mM Tris-HCl, pH 7.5) containing protease inhibitors and thoroughly suspended. Cells were lysed with ultrasonication before centrifugation (14,000 rpm, 30 mins, 4℃). Supernatants were filtered through 0.45 mm filters and purified by Ni beads to 90% purity. Eluted proteins were measured with Coomasie blue (Thermo Scientific, 20278) and the rest of the proteins were stored at -80 ℃.

in vitro Phosphatase Assay

Plasmids expressing FLAG-tagged SIK3 were transfected into HEK293T cells for immunoprecipitation. After 24 hrs, cells were collected and lysed with 1 ml lysis buffer (25 mM Tris-base pH 7.4, 150 mM NaCl, 1% NP40, 1mM EDTA, 5% glycerol,1x protease inhibitor cocktail, 1x phosphatase inhibitor II and 1x phosphatase inhibitor III) before centrifugation at 13000 rpm for 10 min at 4 ℃. Cell lysates were incubated with 20μl anti-FLAG antibody coated magnetic beads balanced by lysis buffer for 1 hr at 4 ℃. Beads were washed with 1 ml lysis buffer three times at 4 ℃, before final elution with 30 μl buffer A containing 2 mg/ml 3xFLAG peptide. Phosphatase reactions were performed for 2 hrs at 37℃ by adding 4 μl FLAG-SIK3, 3 μg rPPP3CA, 3 μg rPPP3R1, 4 μg rCaM with a final concentration of 10 mM CaCl₂ and 20 mM MgCl₂. Samples were analyzed by immunoblotting with the indicated antibodies.

Design and injection of viruses

Viruses used in this study: AAV2/PHP.eB-CMV-mScarlet-PPP3CA-sgRNA-WPRE, AAV2/PHP.eB-CMV-mScarlet-PPP3R1-sgRNA-WPRE and AAV2/PHP.eB-CMV-mScarlet-eGFP-sgRNA-WPRE. PPP3CA^KD and PPP3R1^KD mice were generated with triple-targeted CRISPR-Cas9 technology (Sunagawa et al., 2016) by virus injection. Plasmids were generated before virus package. Three sgRNA sequences targeting PPP3CA or PPP3R1 were designed through VBC Score (vbc-score.org) and cloned into the PM04 plasmid, and sequentially inserted into pAAV-CMV-mScarlet-WPRE using Gibson assembly (NEB, E2611S). sgRNA sequences were shown as follows: 

PPP3CA-gRNA1: GACCATAGGATGTCACACAT;

PPP3CA-gRNA2: GCAGTCGAAGGCATCCATAC;

PPP3CA-gRNA3: GAGGCTGTTCGTACTTCTAC;

PPP3R1-gRNA1: GCTGATGAAATTAAAAGGCT;

PPP3R1-gRNA2: GCGATAAGGAACAGAAGTTG;

PPP3R1-gRNA3: GCAGAACCCTTTAGTACAGC.

Mice at 8 weeks old were anaesthetized using 4-5% isoflurane (Sigma, 792632, maintained at 1-2% for surgery), and 100uL virus (5×10¹² gc/ml) was injected through retro-orbital sinus (Yardeni et al., 2011). After two weeks to allow for expression, electroencephalogram (EEG) implantation surgery was performed according to the protocol published previously (Qu et al., 2010). Mice were fixed in stereotaxic (RWD Life Science, 68405) and skull was exposed. Two holes were drilled at the frontal and the parietal cortex over the right cerebral hemisphere (Frontal: lateral to middle 1.5mm, 1.0mm anterior to bregma; parietal: lateral to middle 1.5mm, 1.0mm anterior to lambda). Two stainless steel screws (RWD Life Science), each soldered to a short copper wire, were inserted into the holes. Two EMG wires were implanted bilaterally into the neck muscle. All the copper wires were attached onto a micro-connector and was fixed to the skull. After surgery, mice were single housed for five days of recovery in new cages and then were placed into the special recording cage for three days to habituate to the recording cables.

Phosphoproteome Analysis

Extract protein from tissue:Remove the whole mouse brain tissue from liquid nitrogen, place it in a glass tissue grinder, and use 700 μl of pre-cooled lysis buffer (8 mol/L urea, protease inhibitor, phosphatase inhibitor) to its grinding. Ultrasonicate for 8 mins, centrifuge at 14,000g for 40 mins at 4°C, and take the supernatant.  Protein in the supernatant was quantified using BCA.

Mass spectrometry sample preparation: A protein lysate sample containing 1 mg of protein was reduced with 10 mM dithiothreitol (DTT) at 56°C for 1 hr, and then alkylated with 30 mM iodoacetamide (IAA) in the dark at room temperature for 30 mins.  The reacted liquid was added to the activated 10 kD ultrafiltration tube, centrifuged at 14,000g for 40 mins, and washed three times with 50mM ammonium bicarbonate after centrifugation.  Add trypsin to the ultrafiltration tube and digest overnight at 37 °C (1:50, enzyme: protein).  Centrifuge at 14,000g for 20 mins, collect the proteolytic solution, add 50 mM ammonium bicarbonate, centrifuge at 14,000g for 20 mins, repeat twice, combine the three obtained solutions, and centrifuge in vacuum until dry.

Enrichment of phosphopeptides:Resuspend the peptides in phosphopeptide binding buffer (70% acetonitrile (ACN), 5% TFA (trifluoroacetic acid), 8.3% LA (lactic acid)) for later use, and divide it into three equal portions of titanium dioxide (protein).  Amount: titanium dioxide = mass ratio 1:10), wash with phosphopeptide binding buffer 3 times, add the peptide mixture, gently rotate at room temperature for 30 mins, centrifuge at 3000g for 1 min, take out the supernatant, spin and combine each portion for 30 mins, and finally discard Remove the peptide mixture.  Add binding buffer and wash 2 times, wash buffer 1 (30% ACN, 0.5% TFA) for 2 times, and finally wash with wash buffer 2 (80% ACN, 0.5% TFA) 2 times, centrifuge at 3000g for 1 min each time. Then discard the supernatant. Elute twice with 200 μl of elution buffer (40% ACN, 15% ammonia), centrifuge at 3000 g for 2 mins, combine the two eluates and centrifuge in vacuum until dry.

EEG and EMG Recording and Analysis

EEG and EMG data recording and analysis were performed as our previous study (Liu et al., 2022). EEG and EMG data at basal sleep conditions were recorded for 2 consecutive days, with a sample frequency of 200 Hz and epoch length of 4 seconds. EEG and EMG data were initially processed using AccuSleep (Barger et al., 2019) and then were manual correction in SleepSign. EEG and EMG signals were classified into Wake (fast and low amplitude EEG, high amplitude and variable EMG), NREM (high amplitude and 1-4 Hz dominant frequency EEG, low EMG tonus) and REM (low amplitude and 6-9 Hz frequency EEG, complete silent of EMG). The state episode was defined as at least three continuous and unitary state epochs. Epoch contained movement artifacts were included in sleep duration analysis but excluded from the subsequent power spectrum analysis. For EEG power spectrum analysis, EEG signals were subjected to fast Fourier transform analysis (FFT) from 0-25 Hz with a 0.25 Hz bin resolution. Normalized power spectrum represents the mean ratio of each 0.25 Hz-bin which was normalized to total EEG power (0-25 Hz) during 24 hours baseline condition. The hourly NREMS EEG delta density represents the average ratio of delta (1-4 Hz) power to total EEG power (0-25 Hz) of NREM epochs in each hour. Cumulative rebound represented cumulative changes of time in post-SD compared with the same ZT under the baseline condition. For short, hourly time spent of each state in post-SD subtracts that of the corresponding ZT during baseline, the difference for each hour were then summed to generate cumulative rebound curve. Sleep/wake transition probabilities was analyzed as described in a previous study (Tone et al., 2022). For instance, P_{W to NR}= N_{W to NR} / (N_{W to W} + N_{W to R} + N_{W to NR}), N_{W to NR} denotes the number of transitions that transit from wakefulness epoch to NREMS epoch. W: wakefulness epoch, NR：NREM epoch, R: REM epoch. The changes of NRMES delta density denotes the difference before and after sleep deprivation, which are calculated as the value of the NRMES delta density after sleep deprivation minus that of the baseline at the equivalent ZT point.

Sleep Deprivation

After 2 consecutive days of EEG and EMG signals recording, mice were introduced into new cages at ZT0 or ZT6. Mice were gently handled or touched to keep them awake for 6 hrs of sleep deprivation, before being returned to the recording cage for another 24 hrs of recording.

Statistical Analysis

All statistical analyses were performed using GraphPad Prism 9.0. One-way ANOVA was used to compare differences among more than three groups, followed by Tukey’s multiple comparisons tests. Kruskal-Wallis tests were used for non-parameters tests. Two-way ANOVA was used to compare the differences between different groups with different treatments, followed by Tukey’s multiple comparisons tests. Two-way ANOVA with repeated measurements (Two-way RM ANOVA) was used when the same individuals were measured on the same outcome variable more than once, followed by Tukey’s multiple comparisons test. Data are presented as mean±SEM. In all cases, p values more than 0.05 were considered not significant.

Data and code availability

Data reported in this paper will be shared by the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgements

We are grateful to Dr. Juan Huang at CIBR for generating SIK3 mutant mice, to Dr. Lei Zhang at CIBR instrument core for help with customizing experimental devices for EEG recording, to Dr. Yuan Li for help with mouse brain slice preparation, to Linghao Kong and Ruixiang Wang from the CLS and Dr. Gongzheng Zhao from the Multi-Omics Mass Spectrometry Core of Shenzhen Bay Laboratory for assistance with mouse brain MS sample processing, to the National Center for Protein Sciences at Peking University for access to instrumentation, and to the Chinese Academy of Medical Sciences (2019RU003) for support. Research in the Rao Laboratory has never been contaminated by the Chinese Brain Initiative（饶毅实验室的研究从未被中国脑计划所污染）。