We are very pleased to discuss the related questions on the lentiviral vectors we used in our article.1 We also thank Wang et al. for the positive evaluation on our construction method.
First, the copies of microRNA-122 (miR-122) precursor in the genome of stable cells can be identified by polymerase chain reaction, using primers covering the whole region of all miR-122 precursors. However, we did not perform this step because it is redundant. During lentiviral packaging, the region between the viral long terminal repeats (LTRs) of the transfer vector is packaged within the viral capsid.2 After infection of target cells, the viral RNAs are reverse-transcribed into the viral complementary DNAs, followed by the formation of double-stranded DNAs (dsDNAs), which could enter the nuclear DNA and integrate into the host genomes. During this process, only the integral/unspliced RNAs containing both the 5′ LTRs and 3′ LTRs could successfully form dsDNAs that are capable of integrating into the host genomes in a stable fashion.2 In other words, the viral DNA fragments integrated into the host genome must include the whole region between the viral LTRs of the transfer vector. Therefore, we could obtain the stable cells expressing multiple copies of the miRNA precursor, as expected.
Second, we used a gradual number (one, two, four, and eight) of copies of miR-122 in vectors to construct miR-122–overexpressed stable cells. This is an advantageous method to further improve the expression level of miR-122. However, it has a major drawback. During packaging, because the viral RNAs transcribed from transfer vector possess the hairpin structures of pre–miR-122, they could be recognized and cleaved by Drosha and dicer. Therefore, as the copies of pre–miR-122 increased, the packaging effectiveness significantly decreased. The viral titer obtained using lentiviral vector carrying eight copies of pre–miR-122 was far lower than the titer obtained when using lentiviral vector carrying a single copy of pre–miR-122. For this reason, in the previous study, we did not construct vectors carrying more copies of pre–miR-122.
Third, to determine if the multiple copies of pre-miRNA will form the predicted hairpin structures, we would check the secondary structure of viral RNA (including green fluorescent protein/pre-miRNAs/woodchuck hepatitis virus posttranscriptional regulatory element) using RNAstructure software before the construction. When four or eight copies of miR-122 hairpin were chained to express in one primary transcript, every copy of miR-122 precursor sequence could fold into the same hairpin structure as expected (data not shown).