Transcriptome analyses of β-thalassemia -28 (A>G) mutation using isogenic cell models generated by CRISPR/Cas9 and asymmetric single-stranded oligodeoxynucleotides (assODN)

β-thalassemia, caused by mutations in the human hemoglobin (HBB) gene, is one of the most common genetic diseases in the world. HBB –28 (A>G) mutation is one of the five most common mutations in China patients with β-thalassemia. However, few studies have been conducted to understand how this mutation affects the expression of pathogenesis related genes including globin genes due to limited homologous clinical materials. Therefore, we first developed an efficient technique using CRISPR/Cas9 combined with asymmetric single-stranded oligodeoxynucleotides (assODN) to generate a K562 cell model of HBB −28 (A>G) named K562−28 (A>G). Then, we systematically analyzed the differences between K562−28 (A>G) and K562 at the transcriptome level by high-throughput RNA-seq pre- and post-erythrogenic differentiation. We found that HBB −28 (A>G) mutation not only disturbed the transcription of HBB but also decreased the expression of HBG, which may further aggravate the thalassemia phenotype and partially explain the severer clinical outcome of β-thalassemia patients with HBB −28 (A>G) mutation. Moreover, we found K562−28 (A>G) cell line is more sensitive to hypoxia and showed a defective erythrogenic program compared with K562 before differentiation. In agreement, p38MAPK and ERK pathway are hyperactivated in K562−28 (A>G) after differentiation. Importantly, all above mentioned abnormalities in K562−28 (A>G) were reversed after correction of this mutation with CRISPR/Cas and assODN, confirming the specificity of these phenotypes. Overall, this is the first time to analyze the effects of the HBB - 28 (A>G) mutation at whole-transcriptome level based on isogenic cell lines, providing a landscape for further investigation of the mechanism of β-thalassemia with HBB −28 (A>G) mutation.


157
(A>G) and K562 cells, indicating high reproducibility of our data ( Fig. 2A). 158 Overall, the gene expression levels between K562 -28 (A>G) and wt cell are 159 similar transcriptome-wide, suggesting that the mutation may affect (Fig. 2B). 160 We conducted analysis of differentially expressed genes (DEG) and found 161 120 and 524 genes were consistently upregulated and downregulated in K562 -162 28 (A>G) compared to K562 ( Figure 2C). To further explore the affected 163 underlying biological functions, we conducted GO term, KEGG and 164 Reactome analysis using those DEGs and found pathways of (cellular) 165 response to hypoxia and response to (decreased) oxygen level were 166 upregulated in K562 -28 (A>G) mutant cell line ( Figure 2D). In consistent, 167 hypoxia related genes such as HMOX1, BMP7, GATA6, ESAM, RYR2 were 168 upregulated in K562 -28 (A>G) ( Figure 2C). Interestingly, PI3K-Akt signaling 169 pathway that is important for erythrocyte differentiation [18] , was 170 downregulated in K562 -28 (A>G) (Fig. 2D). To further explore the core 171 regulators, we performed interaction assay to predict transcription factors 172 (TFs) that target upregulated genes in K562 -28 (A>G) . In agreement with 173 previous results, we observed GATA family, HOXD10 and SPIC, which are 174 well-known regulators of erythroid differentiation and hypoxia [ref]，were 175 the core regulators for upregulated genes in K562 -28 (A>G) (Fig. 2E). The regulators and their corresponsive target genes were listed and genes related 177 to hypoxia response were labelled in red (Fig. 2F). Taken together, these data 178 suggested hypoxia response was upregulated in K562 -28 (A>G) and the core 179 regulators were GATA family, HOXD10 and SPIC. been reported to be activated during erythroid differentiation [18][19][20][21] . In DNA elements [22] . In b-thalassemia patients, HBG may be upregulated to 204 compensate the loss of HBB. To study the compensatory gene expression of 205 globin genes in K562 -28 (A>G) after differentiation, the expression of HBB, HBA and HBG was analyzed by RNA-sequencing in isogenic cell lines of K562 -28 207 (A>G) and K562. As expected, Integrative Genomics Viewer(IGV)analysis 208 showed that HBB expression was induced in K562 but undetectable in K562 -209 28 (A>G) after differentiation (Fig. 3D). In contrast, expression of other globin 210 genes was induced and detected in K562 -28 (A>G) after differentiation (Fig. 3E). 211 we noticed that fold induction rate of HBA1, HBA2, HBE1 and HBZ were 212 similar or slightly increased in K562 -28 (A>G) -Dif when compared to those in including GATA1, BGL3 and NFE2 [19,[23][24][25] , were dramatically upregulated 226 in K562 after differentiation, while those increasement were largely 227 attenuated in K562 -28 (A>G) . In contrast, negative regulator BCL11A was 228 upregulated in K562 -28 (A>G) before differentiation (Fig. 3I). Those data 229 suggested the erythrogenic differentiation was overall normal in K562 -28 (A>G) ,   cell line as K562 -28(A>G)cor . We analyzed the correlation of RNA-seq data 269 between K562 -28(A>G)cor , K562 and K562 -28 (A>G) and found expression 270 profile of K562 -28(A>G)cor was closer to K562 rather than its precursor K562 -271 28 (A>G) (Fig. 5B). DEGs of isogenic cell lines were showed in the form of 272 heat map and results indicated that upregulated and downregulated genes in  Our results showed HBB -28 (A>G) mutation prevented the transcription of 291 HBB gene. Analysis of enriched pathways suggested PI3K pathway, as well 292 as JAK-STAT pathway, which play important roles in the erythroid 293 differentiation, were disrupted in K562 -28 (A>G) before erythroid differentiation.

294
The PI3K-Akt signaling pathway is a significant pathway that control many cellular processes known as cell division, autophagy, survival, and 296 differentiation [ref]. Moreover, the mutation activated the hypoxia pathway 297 in undifferentiated K562 -28 (A>G) . Many clinical manifestations observed in 298 b-thalassemia is attributed to the chronic hypoxic environment due to 299 pathologic erythrocyte production, and our data suggest hematopoietic 300 precursors may also subject to oxidative stress before differentiation.

302
To induce erythroid differentiation, we chose the glutamine-minus medium 303 with sodium butyrate, as hemoglobin synthesis was markedly induced using 304 this condition with a differentiation efficiency of 11%~19% in K562 [26,27] .

305
Consistent with previous reports, the differentiation efficiency of K562 in our 306 study was nearly 12% (data not shown), indicating that our erythroid 307 differentiation is effective. In agreement, the MAPK and ERK pathway was 308 activated in both K562 and K562 -28 (A>G) (Fig. 2), a finding consistent with 309 observations in previous studies [20,28] . Interestingly, PI3K-Akt signaling 310 pathway was activated in K562 -28 (A>G) after induction, suggesting the 311 defective PI3K pathway may be caused by lack of activators in 312 undifferentiated K562 -28 (A>G) . Other pathways, such as cell adhesion, 313 pluripotency of stem cells, platelet activation and Notch pathway, were also 314 co-activated in differentiated K562 and K562 -28 (A>G) samples, indicating 315 mutation of HBB -28 (G>A) didn't block the pathways required for 316 differentiation. Nevertheless, in consistent with data from undifferentiated 317 K562 -28 (A>G) , oxygen related pathways were downregulated in differentiated 318 K562 -28 (A>G) (Fig. 4). In both undifferentiated and differentiated conditions, 319 SPIC and GATA families are predicted as core regulators ( Fig.2 and Fig.4).

320
The GATA family of transcription factors (GATA1-6) are essential for normal 321 hematopoiesis and a multitude of other developmental processes [19,29] . GATA-322 1 regulates terminal differentiation and function of erythroid which activates 323 or represses erythroid-specific gene, such as b-globin locus-binding protein, 324 and it might regulate the switch of fetal to adult haemoglobin in human [30] .
Interestingly, increased expression of GATA1 was largely attenuated in K562 -326 28 (A>G) during erythroid differentiation (Fig. 3), which may play a role for 327 dysregulated oxygen related pathways.

329
As improving the levels of HBG in adults could partially reverse the severity 330 of symptoms in sickle disease and β-thalassemia, it is important to understand 331 the coordinated regulation between HBB and HBG [31][32][33][34] . In this study, we 332 noticed that fold induction of HBG was decreased in K562 -28 (A>G) . ZBTB7A 333 and BCL11A were two major repressors of HBG by directly bound the HBG 334 gene promoters [24,[35][36][37] . Expression of ZBTB7A was decreased during 335 differentiation in K562, consistent with increased expression of HBG. 336 However, the overall expression level was higher in K562 than that in K562 -  Collectively, our data suggest the expression of HBG was not only regulated 342 by ZBTB7A and BCL11A, but may also be regulated by GATA proteins that 343 also regulate HBB. However, the interaction requires further investigations.   We would like to thank Yanmei Deng, Jiaying Zhang and Wenwen Yao for 365 helping single cell identification. We thank the Genome Synthesis and Editing

366
Platform of the China National GeneBank for providing support on gene 367 synthesis. We thank BGI colleagues for helping to output the high-quality data.  according to the protocol provided by Zhang F's protocol [38] .

387
Design of ss ODN repair templates :The 127-nt asymmetric ss ODN repair  differentiation, and then collected [39] . And the control cells were cultured in 408 RPMI 1640 medium with 10% FBS and P/S antibiotics for 7 week at the same 409 time. There are exceeding 2 million 100 bp pair-ended reads in each sample. Prior 446 to assembly, Raw reads were filtered by SOAPnuke [40] with the parameters 447 "-l 15 -q 0.2 -n 0.05". Reads were mapped to the human reference genome 448 using HISAT2 [41] , which included both the genome sequences (GRCh38. The authors declare that they have no competing interest.   Notch signaling pathway cell chemotaxis Cell junction organization regulation of growth Focal adhesion extracellular structure organization angiogenesis response to wounding PI3K−Akt signaling pathway Extracellular matrix organization regulation of cell projection organization negative regulation of cell projection organization ECM−receptor interaction apoptotic signaling pathway regulation of leukocyte migration positive regulation of cell motility positive regulation of GTPase activity response to BMP cellular response to BMP stimulus negative regulation of blood vessel diameter regulation of extrinsic apoptotic signaling pathway Hemostasis cellular response to oxygen levels negative regulation of apoptotic signaling pathway cellular response to decreased oxygen levels cellular response to hypoxia response to oxygen levels response to decreased oxygen levels positive regulation of cardiocyte differentiation response to hypoxia regulation of blood vessel size regulation of blood vessel diameter    (740) response to chemokine Cytokine−cytokine receptor interaction negative regulation of platelet aggregation Erythrocytes take up carbon dioxide and release oxygen O2/CO2 exchange in erythrocytes regulation of sodium ion transmembrane transporter activity negative regulation of platelet activation regulation of cell morphogenesis regulation of blood pressure gas transport NOTCH4 Intracellular Domain Regulates Transcription positive regulation of cell growth regulation of sodium ion transport regulation of angiogenesis TNF signaling pathway positive regulation of leukocyte activation cellular response to metal ion Cell adhesion molecules (CAMs) regulation of Wnt signaling pathway Signaling by VEGF oxygen transport negative regulation of myeloid leukocyte mediated immunity inflammatory response to wounding negative regulation of hemopoiesis Iron uptake and transport negative regulation of blood coagulation negative regulation of hemostasis negative regulation of leukocyte differentiation response to interleukin−7 cellular response to interleukin−7 interleukin−7−mediated signaling pathway negative regulation of leukocyte activation myeloid leukocyte cytokine production negative regulation of blood circulation regulation of apoptotic process involved in development regulation of apoptotic process involved in morphogenesis    BMP7  EGR1  KCND2  MET  IGF1  GNG7  ITGA7  FLT4  COL6A1  CD99  FLT3LG  COL6A2  TNXB  ITGB4  NGFR  KIT  COL9A2  INSR  ITGB3  PCK1  PPP2R3B  IL3RA  IL4R  COL4A2  CD19  JAK3