High-throughput mapping of single-neuron projection and molecular features by retrograde barcoded labelling [scRNA-seq]

Deciphering patterns of connectivity between neurons in the mammalian brain is a critical step toward understanding brain function. Conventional imaging based neuroanatomical tracing methods identify area-to-area or sparse neuron-to-neuron connectivity patterns, but with extremely limited throughput. Recently developed barcode-based connectomics methods can efficiently map large numbers of single-neuron projections, but linking these data to single-cell transcriptomics remains a challenge. Here, we established a retro-AAV barcode-based multiplexed tracing method called MERGE-seq (Multiplexed projection neuRons retroGrade barcodE sequencing), which is capable of simultaneously characterizing the projectome and transcriptome at the single neuron level. We uncovered dedicated and collateral projection patterns of ventromedial prefrontal cortex (vmPFC) neurons to five downstream targets (AI, DMS, BLA, MD and LH). We found that projection-defined vmPFC neurons are molecularly heterogeneous, which are composed of different neuronal subtypes. We further identified transcriptional signatures of various dedicated and bifurcated vmPFC neurons, and verified Pou3f1 as the marker gene of neurons sending collateral axons to DMS and LH. Finally, we fitted our single-neuron connectome/transcriptome data into a machine learning-based model and revealed groups of genes that were predictive of certain projection pattern. In summary, we have developed a new multiplexed technique whose paired connectome and gene expression data can help reveal organizational principles that form neural circuits and process information. Male adult C57BL/6 mice (8 weeks of age) were anesthetized intraperitoneally using pentobarbital sodium (10 mg/mL, 120 mg/kg b.w.) and unilaterally injected with rAAV-EGFP-BARCODE virus into five projection targets simultaneously. Barcode 0 sequences representing the AI target is: CTGCACCGACGCATT; barcode 1 sequences representing the DMS target is: GAAGGCACAGACTTT; barcode 2 sequences representing the MD target is: GTTGGCTGCAATCCA; barcode 3 sequences representing the BLA target is: AAGACGCCGTCGCAA; barcode 4 sequences representing the LH target is: TATTCGGAGGACGAC. 3 mice were performed scRNA-seq without FAC-sorting, cell pellets were resuspended and 48,000 cells were loaded into 3 lanes to perform 10X Genomics sequencingperformed scRNA. 3 mice with sorting and pooled together for scRNA-seq. Chromium Single Cell 3' Reagent Kits (v3) were used for transcriptome library preparation (10X Genomics). Libraries were sequenced on an Illumina Novaseq 6000 system.破译哺乳动物大脑神经元之间的连接模式，是理解大脑功能的关键一环。传统基于成像的神经解剖学追踪方法可识别脑区-脑区或稀疏的神经元-神经元连接模式，但通量极低。近年来开发的基于条形码的连接组学方法能够高效绘制大量单个神经元的投射模式，但将此类数据与单细胞转录组学关联仍是一项挑战。 本研究建立了一种基于逆行腺相关病毒（retro-AAV）条形码的多重追踪方法，命名为MERGE-seq（Multiplexed projection neuRons retroGrade barcodE sequencing，多重投射神经元逆行条形码测序），可在单个神经元水平上同时解析投射组与转录组。我们揭示了腹内侧前额叶皮层（ventromedial prefrontal cortex, vmPFC）神经元向5个下游靶点（AI、DMS、BLA、MD和LH）的专一投射与侧支投射模式。研究发现，以投射特征定义的vmPFC神经元具有分子异质性，由不同的神经元亚型构成。我们进一步鉴定了各类专一投射和分叉投射vmPFC神经元的转录特征，并验证Pou3f1可作为向DMS和LH发出侧支轴突的神经元的标记基因。最后，我们将单个神经元连接组/转录组数据拟合至基于机器学习的模型中，揭示了可预测特定投射模式的基因集合。综上，本研究开发了一种新型多重技术，其配对的连接组与基因表达数据有助于揭示构成神经环路并处理信息的组织原则。 本研究选用8周龄的成年雄性C57BL/6小鼠，通过腹腔注射戊巴比妥钠（10 mg/mL，120 mg/kg体质量）进行麻醉，同时向5个投射靶点单侧注射rAAV-EGFP-BARCODE病毒。代表AI靶点的条形码0序列为：CTGCACCGACGCATT；代表DMS靶点的条形码1序列为：GAAGGCACAGACTTT；代表MD靶点的条形码2序列为：GTTGGCTGCAATCCA；代表BLA靶点的条形码3序列为：AAGACGCCGTCGCAA；代表LH靶点的条形码4序列为：TATTCGGAGGACGAC。 其中3只小鼠未进行荧光激活细胞分选（FAC-sorting），直接制备样本进行单细胞RNA测序（scRNA-seq）：将细胞沉淀重悬后，取48000个细胞分配至3个测序通道，用于10X Genomics测序。另有3只小鼠经分选后混合样本进行scRNA-seq。转录组文库制备采用Chromium单细胞3'试剂试剂盒（v3，10X Genomics）。文库在Illumina Novaseq 6000系统上完成测序。