Figure 3 - N-glycan characterization of miRNA KO/KD cell lines. Statistically significant difference (*, p<0.05) antennary structures among secreted proteins from the mock (px458 F6 and px458 H12) and KD/KO (miR-128 C3, miR-128 E7, miR-34c D12, and miR-34c F3) cell lines based on N-glycosylation profiling. Details are found in Supplementary Figure 1.
In a qPCR analysis of glycosylation gene expression in the culture (Figure 5), it seems that miR-128 KD may regulate Mgat4b and Mgat5 by increasing their expression levels. In addition, miR-34c KD may regulate the expression levels of Mgat4a and Mgat4b, and possible also Mgat5. On the other hand, SLC35A3 and SLC35A5 seems not to be directly targeted by mir-128 and miR-34c, as expression levels of SLC35A3 and SLC35A5 in the corresponding miR KD cell lines are similar to the mock cell line.

Construction and validation of miRNA over-expression (OE) cell pools

In order to test more miRNA targets, we constructed seven cell pools (Table 1), each overexpressing one miRNA. We tested the expression with qPCR, and all miRNA OE cell pools show relative overexpression of corresponding miRNAs (Figure 4A). The levels of the relative fold change are correlated well with the level of corresponding of endogenous miRNAs expressed in the original CHO-S cell lines. The miRNAs that has lower expression levels in original CHO-S cell lines, for example mir-122 and miR30a, shows higher fold change when overexpressing the corresponding miRNAs.

Glycoprofiling of OE cell pools

The next step was to explore the phenotype of the miRNA overexpression. We did this by glycoprofiling the secreted protein from the seven pools and comparing them to a mock (Figure 4B). It is clear, that only overexpressing miR-30b and miR-449a gave detectable changes in N-glycan profiles, namely reduced levels of tri- and tetra- antennary structures, and increase mono- and bi- antennary structures. This indicates that miR-30b and miR-449a may also target genes that regulate GlcNAc transferase reactions for addition GlcNac onto β 1,4 (Mgat4a, Mgat4b) and β 1,6 branches (Mgat 5), and/or target UPD-GlcNAc tansporters (SLC35A3, SLC35A4, SLC35A5), or possibly enzymes working earlier in the N-glycan pathway. According to qPCR results of possible target genes (Figure 6), we found that miR-30b levels correlate inversely with the levels of Mgat4a, SLC35A3 and SLC35A5. Thus, miR-30b seems to target Mgat4a, SLC35A3 and SLC35A5 mRNAs. Therefore, miR-30b may change antennary structures of N-glycan by simultaneously regulate GlcNAc transferase reactions and the availability of UDP-GlcNAc inside Golgi apparatus. miR-449a, on the other hand, may only target Mgat4a and reduce the levels of β 1,4 branched N-glycans.