Date Published: September 5, 2013
Publisher: Public Library of Science
Author(s): Amelia M. Gallaher, Sudipto Das, Zhen Xiao, Thorkell Andresson, Philippe Kieffer-Kwon, Christine Happel, Joseph Ziegelbauer, Dirk P. Dittmer.
Kaposi’s sarcoma (KS) is caused by infection with Kaposi’s sarcoma-associated herpesvirus (KSHV). The virus expresses unique microRNAs (miRNAs), but the targets and functions of these miRNAs are not completely understood. In order to identify human targets of viral miRNAs, we measured protein expression changes caused by multiple KSHV miRNAs using pulsed stable labeling with amino acids in cell culture (pSILAC) in primary endothelial cells. This led to the identification of multiple human genes that are repressed at the protein level, but not at the miRNA level. Further analysis also identified that KSHV miRNAs can modulate activity or expression of upstream regulatory factors, resulting in suppressed activation of a protein involved in leukocyte recruitment (ICAM1) following lysophosphatidic acid treatment, as well as up-regulation of a pro-angiogenic protein (HIF1α), and up-regulation of a protein involved in stimulating angiogenesis (HMOX1). This study aids in our understanding of miRNA mechanisms of repression and miRNA contributions to viral pathogenesis.
At our current understanding, the herpesvirus family is the only viral family expressing multiple miRNAs. Kaposi’s sarcoma-associated herpesvirus (human herpesvirus 8) expresses 12 pre-miRNAs , , , . These miRNAs are encoded in the latency locus of the KSHV genome and all KSHV miRNAs are expressed during latency. This discovery presented the possibility that KSHV expresses miRNAs to modulate host gene expression by a mechanism that would avoid generating additional viral proteins, which could be detected by the host immune system.
In order to identify the target genes repressed by KSHV miRNAs, we measured the effects of KSHV miRNAs on protein expression by introducing viral miRNAs into uninfected primary endothelial cells (HUVEC). Primary cells were transfected with either a control non-targeting miRNA mimic or a combination of sixteen KSHV miRNA mimics (Figure S1). We transfected miRNA mimics in the absence of viral infection to ensure that the repression of newly synthesized proteins was not an indirect result of infection or viral protein expression. After transfection, cells were grown in media supplemented with two distinct mixtures of stable medium-heavy or heavy amino acids. Using this pulsed labeling approach we measured newly translated proteins from the two conditions (Figure 1A) with mass spectrometry. The relative amount of mature miRNAs in RNA-induced silencing complexes was probed using Argonaute 2 immunoprecipitations, followed by reverse-transcription and quantitative PCR (normalized to human miR-21, which does not change upon KSHV infection ). Transfected miRNA mimics were indeed associated with the RNA-induced silencing complex (RISC) (Figure 1B) using this assay. Before mass spectrometry analysis, the samples were analyzed for expected repression of previously identified miRNA targets (BCLAF1, TWEAKR) using quantitative Western blotting (Figure 1C). Cells from the two labeling conditions were combined in a 1∶1 ratio, and proteins were extracted, fractionated and analyzed by tandem mass spectrometry to calculate relative changes in newly translated proteins due to the presence of KSHV miRNAs. Fractionation of the protein samples was utilized to improve coverage of a wide variety of proteins and to better detect less abundant proteins. Both biological replicates were analyzed by mass spectrometry twice, yielding two technical replicates for two biological replicates. Mass spectrometry data was filtered for proteins detected by at least two peptide pairs (medium-heavy and heavy) per replicate and detected in both biological replicates. There were 1276 proteins that met these stringent quality control filters (Figure 1D) and the most down-regulated proteins in cells containing KSHV miRNAs are shown in Figure 1E. It was noteworthy that thrombospondin (THBS1), the first identified target of KSHV miRNAs was strongly inhibited at the protein level .
In order to understand miRNA functions, it is critical to identify their targets, so we can increase our knowledge of cellular pathways that are important for infection and pathogenesis. Genome-wide studies have been conducted analyzing the Argonaute-associated mRNAs (CLIP assays) in B cells , , , and the microarray and proteomic screening for miRNA-induced gene expression changes in primary endothelial cells from this report represent a complimentary dataset for elucidating viral miRNA functions. Indeed, integration of miRNA targets from CLIP methods and other expression studies will continue to be useful for identifying miRNA target sites, as well as >those CLIP hits that are repressed at the mRNA and/or protein level. Compared with other approaches to discover miRNA targets, current mass spectrometry methods are able to query a lower number of gene products. Despite this limitation, this current study has identified repression of multiple novel and previously validated miRNA targets (THBS1, GRB2). Additionally, gene expression studies can reveal direct and indirect miRNA targets, both of which are important for virus-host interactions. By inspecting gene expression changes at both the mRNA and protein level, we have demonstrated that multiple miRNA targets are likely missed using microarrays since the miRNA target may only be repressed at the level of translation. This finding is relevant given the conflicting reports about the predominant mechanism and order of repression mechanisms  that are utilized by miRNAs to modulate gene expression, whether that be mRNA level repression  or translation inhibition , . In this study, validated miRNA targets AKAP9, STAT3, and GRB2 proteins were significantly repressed, but microarray results indicated mRNA levels were not reduced in the presence of KSHV miRNA mimics. The protein SH3-domain GRB2-like endophilin B1 (SH3GLB1) was the second most inhibited protein, but the mRNA levels were relatively unchanged (log2 0.03). Interestingly, previous reports have shown that SH3GLB1 functions as a tumor suppressor and pro-apoptotic factor , . Given our findings, this proteomic method is clearly an important start to discover novel miRNA targets. Furthermore, we have also shown novel functions of viral miRNAs involved in cellular pathways important to KSHV pathogenesis, including ICAM1 repression, HMOX1 up-regulation and HIF1α up-regulation.