Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012 Jul;22(7):1243-54.
doi: 10.1101/gr.132514.111. Epub 2012 Mar 28.

Impact of microRNA regulation on variation in human gene expression

Jian Lu  1 , Andrew G Clark
Affiliations

Impact of microRNA regulation on variation in human gene expression

Jian Lu et al. Genome Res. 2012 Jul.

Abstract

MicroRNAs (miRNAs) are endogenously expressed small RNAs that regulate expression of mRNAs at the post-transcriptional level. The consequence of miRNA regulation is hypothesized to reduce the expression variation of target genes. However, it is possible that mutations in miRNAs and target sites cause rewiring of the miRNA regulatory networks resulting in increased variation in gene expression. By examining variation in gene expression patterns in human populations and between human and other primate species, we find that miRNAs have stabilized expression of a small number of target genes during primate evolution. Compared with genes not regulated by miRNAs, however, genes regulated by miRNAs overall have higher expression variation at the population level, and they display greater variation in expression among human ethnic groups or between human and other primate species. By integrating expression data with genotypes determined in the HapMap 3 and the 1000 Genomes Projects, we found that expression variation in miRNAs, genetic variants in miRNA loci, and mutations in miRNA target sites are important sources of elevated expression variation of miRNA target genes. A reasonable case can be made that natural selection is driving this pattern of variation.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Scheme for a canonical miRNA regulation network. The comprehensive interactions between miRNAs and transcription factors (TF) are expected to comprise “wired” genetic networks to regulate the expression of target genes. (A,B) Examples of incoherent and coherent feed-forward loops, respectively. In the incoherent feed-forward loop, the direct regulatory effect of a TF on the target gene is opposed to the indirect regulatory effect through miRNA regulation (A); and in the coherent feed-forward loops, the direct regulatory effect of a TF on the target gene is synergistic to the indirect regulatory effect through miRNA regulation (B). The consequence of miRNA regulatory effects is to reduce the stochastic noise in expression levels of target genes. (C,D) Distributions of expression levels (the x-axis is the number of molecules per cell) of two hypothetical target genes across cells (or individuals). (C) Expression levels of the target gene are tightly regulated due to miRNA targeting; (D) expression levels of a gene not targeted by miRNAs are highly variable across individuals. (E,F) Six sources of genetic variation in the miRNA regulatory networks. (E) The scheme of a miRNA precursor characterized by a hairpin structure. The sequence in black is the mature miRNA (guide strand) and position 2–8 of mature miRNA is the “seed” region (underlined). Perfect pairing between the seed of the mature miRNA and the target site is crucial for miRNA target recognition (F). Mutations associated with a miRNA precursor can be divided into four categories: (1) mutations in the “seed” alter the target recognition; (2) changes in the mature miRNA beyond the “seed” region potentially affect target recognition; (3) changes outside mature miRNAs can affect miRNA biogenesis and hence affect the abundance of the mature miRNA; and (4) changes in promoter regions of the miRNA precursor will cause the abundance of mature miRNA to be variable (E). Mutations in miRNA target pairing regions also affect miRNA binding: (5) Mutations in the seed pairing region of a miRNA will affect the target recognition and hence the expression level of the host genes; and (6) mutations in regions of 3′ UTR beyond seed pairing might affect the accessibility of a miRNA to the target site. In our model, genetic variation in the former four classes is defined as the “trans-regulatory” effect, and variation in the latter two classes is defined as the “cis-regulatory” effect. Both trans- and cis-regulatory effects in the miRNA regulatory networks contribute to the expression variation of the target genes.
Figure 2.
Figure 2.
Expression variability (CV) in the lymphoblastoid cells is significantly higher for genes that are targeted by the coexpressed 287 miRNAs, regardless of target determination methods. (A) The results for data set I (Wang et al. 2009); (TSs) TargetScan with conservation (PCT > 0.8); (TSt) context score (≤0.4 and percentile >85); (Verified) experimentally verified targets taken from TarBase, miR2Disease, and Hafner et al. (2010). (B) The results after combining miRNA target determination methods together. (High-confidence miRNA targets) Only targets determined by TSs (PCT > 0.8); (TSt) (≤0.4 and percentile >85), Pictar, or experimentally verified targets of the 287 coexpressed miRNAs; (Non-targets) transcripts that are not targeted by the 287 coexpressed miRNAs with any of the miRNA target determination methods (including the four aforementioned methods, PITA, and miRanda algorithms). Boxplots of CVs for targets (gray) and non-target transcripts (white) are presented.
Figure 3.
Figure 3.
The magnitude of the miRNA regulation effect is positively associated with gene expression variability. In each plot, the x-axis is the number of target sites. (A,B) The number of the putative miRNA target sites predicted by TSt (context score ≤ 0.3) and the median CV for the transcripts in data sets I and II. The number of distinct miRNA target sites that are harbored in an mRNA 3′ UTR varies from 0 to 14 (a small number of mRNAs have >14 sites and are binned into 14). (C,D) The number of conserved miRNA-interacting sites (based on TSs PCT > 0.8) and the median CV of the genes. The number of conserved target sites located in one transcript is generally smaller than the number of sites predicted by context score, and varies from 0 to 5 (a small number of mRNAs have more than five sites and are binned into five). Only microarray probes that are mapped on the RefSeq genes are used in the two data sets.
Figure 4.
Figure 4.
High-confidence miRNA target genes tend to display greater differential expression between CEU vs. Asian populations (FDR is cut off at 0.001). This pattern holds true either when we incorporate all of the probes or restrict our analysis to the RefSeq transcripts. Similar patterns were observed when we considered both genders, or males alone. High-confidence miRNA targets were considered as miRNA targets, while genes not targeted by coexpressed miRNAs in all of the six methods were considered non-targets (see legend to Fig. 2). (Gray) miRNA targets; (white) non-targets. When we only consider RefSeq transcripts, 3187 of 8660 (36.8%) miRNA target transcripts are differentially expressed, and 818 of 2421 (33.8%) non-target transcripts are differentially expressed (P = 0.007). For all of the expressed transcripts in males, 26.9% (2332 out of 8684) of targets are differentially expressed, and 19.4% (1080 out of 5560) of non-targets are differentially expressed (P < 10−16). For the RefSeq data in males, 25.6% (2217 out of 8660) of targets are differentially expressed, and 22.7% (550 out of 2421) of non-targets are differentially expressed (P < 10−16).
Figure 5.
Figure 5.
Genes that are differentially expressed or under stabilizing selection during primate evolution are significantly enriched in miRNA target genes. (A) The genes that are differentially expressed between human and chimpanzee (deHC). (B) The genes that are differentially expressed between human and macaque (deHR). (C) The genes that are under stabilizing selection. All of the 20,689 Ensembl genes sequenced by Blekhman et al. (2010) are used in A, B, and C. (D) The combinational miRNA target gene determination methods. Only the 16,772 protein-coding genes examined in Blekhman et al. (2010) are used.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. The 1000 Genomes Project Consortium 2010. A map of human genome variation from population-scale sequencing. Nature 467: 1061–1073 - PMC - PubMed
    1. Allen E, Xie Z, Gustafson AM, Sung GH, Spatafora JW, Carrington JC 2004. Evolution of microRNA genes by inverted duplication of target gene sequences in Arabidopsis thaliana. Nat Genet 36: 1282–1290 - PubMed
    1. Ambros V 2003. MicroRNA pathways in flies and worms: Growth, death, fat, stress, and timing. Cell 113: 673–676 - PubMed
    1. Baek D, Villen J, Shin C, Camargo FD, Gygi SP, Bartel DP 2008. The impact of microRNAs on protein output. Nature 455: 64–71 - PMC - PubMed
    1. Bagga S, Bracht J, Hunter S, Massirer K, Holtz J, Eachus R, Pasquinelli AE 2005. Regulation by let-7 and lin-4 miRNAs results in target mRNA degradation. Cell 122: 553–563 - PubMed

Publication types

MeSH terms

LinkOut - more resources