이 장은 원문만 있습니다. 문장별 해석은 앞 장부터 차례로 붙입니다.
3장 · Human evolution: the non-coding revolution — Human variation links accelerated regions to phenotypes
Another way to associate human-specific non-coding regions with traits is to investigate associations between any polymorphisms they harbor and phenotypic variability in humans. The polymorphic sites will be relatively recent mutations that are largely distinct from differences between human and chimpanzee reference genomes. But they may nonetheless shed light on the general function of the accelerated regions in which they occur. For example, gene copy number variants in the human-specific pericentric inversion of chromosome 1 have been associated with human developmental and neurogenetic diseases, and we showed that HARs that are conserved in primates but not across all mammals are associated with schizophrenia. As many more human genomes are sequenced, population genetic variation in and near to HARs will also help researchers to test hypotheses about the evolutionary forces that created and maintain HARs (Box 1, Fig. 3).
Leveraging this approach, Doan et al. sequenced HARs in individuals with autism spectrum disorder (ASD) and unaffected controls. ASD was associated with a significant 43% excess of rare biallelic variants in HARs. The authors showed that 29% of these variants alter enhancer activity in primary mouse neurospheres using a custom MPRA. They further identified rare homozygous mutations with active regulatory marks near neurodevelopmental and disease-associated genes in ASD patients who lack causal coding mutations. Using luciferase reporter assays and transgenic mice, the authors show that one particular ASD-associated HAR variant (HACNS426) that was previously shown to interact with the neuronal morphology gene Cux1 is an enhancer (Fig. 4). The authors report that transgenic mice carrying the mutant allele fused to the CUX1 promoter and to GFP show elevated expression of the reporter gene in the developing brain compared to the wild-type allele. But it is important to remember that transgenic enhancer assays cannot capture quantitative expression differences (see above).
The study of Doan et al. demonstrates the potential for human sequencing studies coupled with deep phenotyping to shed light on HAR function. Using associations between specific HARs and ASD, the authors conclude that many HARs are essential for normal development. Even though the ASD-associated mutations are different from the substitutions that happened during human evolution, they nonetheless help to annotate the function of the genomic region containing a HAR. Polymorphic deletions of HARs could similarly be used to annotate their functions. Doan et al. further hypothesize that human-specific mutations in HARs could have altered social and/or cognitive behavior. To test this hypothesis precisely with medical re-sequencing requires that specific human–chimp differences and combinations thereof be polymorphic in living humans or primates. As targeted sequencing (for example, with molecular inversion probes) enables HARs to be sequenced in millions of individuals, it will become clear how many human–chimp differences are present in modern humans. One caveat is that the genomes of modern humans and other primates differ from those of ancestral populations, so phenotype associations of HAR mutations must be interpreted in the context of the modern genetic background. For this approach to be a useful path to elucidating HAR functions in humans, it will also be important to couple sequencing data with a very wide range of reliably measured phenotypes ranging from diseases to behaviors. This is an exciting and increasingly feasible direction.
Changing things so everything stays the same? Regulatory element turnover and compensatory evolution
To measure the effects of non-coding mutations on expression and phenotypes, we must move towards considering the whole regulatory region of a gene, as well as changes in trans (such as, expression and activity of upstream regulators). One reason for this is that regulatory elements without sufficient sequence similarity to be aligned across species can direct conserved expression patterns, a phenomenon known as ‘enhancer turnover’ (Fig. 3). Another consideration is the fact that regulatory elements interact with each other to drive gene expression, which means that non-coding mutations can amplify or buffer one another even if they are not adjacent. One particular type of compensation across regulatory elements involves ‘shadow enhancers’, which are distinct sequences capable of guiding similar spatiotemporal expression patterns (Fig. 3). This apparently redundant function may ensure robust, precise, and stable regulation. Thus, gene expression evolution cannot be easily predicted from a single human-specific regulatory element without evaluating the complete regulatory landscape of the locus.
Most human-specific regulatory elements have nonetheless been characterized one at a time, or even one mutation at a time. Supporting the idea that this limits our ability to predict expression divergence, Vermunt et al. showed that many loci contain both gains and losses of predicted enhancers (4.9–10.3% across brain regions of mice, macaques, chimpanzees, and humans). For example, the locus containing HAR87 harbors multiple other predicted enhancers with increased or decreased activity during human evolution, measured via chromatin capture coverage (Fig. 2). These other changes could compensate for or otherwise interact with changes in the activity of HAR87. To investigate this possibility, comparative genomics should be combined with functional genomics assays, especially those that do not depend on antibody affinities (for example, chromatin capture, open chromatin).
An emerging approach is to leverage high-throughput genetic screens that can probe combinations of mutations across a locus, for example, via CRISPR genome editing. Using a combination of CRISPR activation and inhibition, as well as knock-outs and knock-ins, it will be possible to study the effect of human–chimpanzee non-coding sequence differences on gene expression in cells or in models such as mice. These studies should investigate all the putative regulatory regions of a gene individually and in combination to truly decipher the net regulatory effects.
Thinking again: what are we missing in the study of genetic bases of human evolution?
The past decade has seen significant progress towards addressing the major hurdles of associating human-specific non-coding elements with traits that make our species unique. Functional genomics has been transformative in terms of annotating and prioritizing HARs, and it has also been used directly to identify novel human-specific regulatory elements that are not divergent enough in sequence to be detectable in genome-wide tests for accelerated sequence evolution. As the mechanisms of gene regulation are increasingly understood, the functional effects of human-specific non-coding mutations are becoming less mysterious. High-throughput functional assays, such as massively parallel reporter assays and CRISPR screens, are increasing the numbers of characterized human-specific regulatory elements by orders of magnitude. Meanwhile, targeted and whole genome sequencing of thousands of people is enabling rare variant association and linkage studies that connect non-coding elements to traits and may even reveal the phenotypes of individuals carrying specific ancestral versus derived haplotypes. Human evolutionary studies will always be challenged by our inability to test specific hypotheses about genetic changes in the correct organismal context (an ancestral human). However, we now have a much broader collection of complementary research tools to address this problem, including on one hand genome editing of stem cell-derived cell lines, organoids, human primary cells, and model organisms, and on the other hand deep phenotyping and sequencing of humans and other primates to capture natural variation.
Despite these exciting advances towards understanding the role of regulatory changes in human evolution, it is important to ask what hurdles still remain. One major barrier is our very limited ability to assay functions of regulatory elements other than enhancers. Genome editing is helping to address this bias by enabling researchers to introduce individual human mutations or to knock out single human-specific elements. By performing these experiments in primate cells, human cells, and model organisms, it will be feasible to characterize putative human-specific insulating and repressing elements in terms of their downstream effects on molecular and cellular phenotypes. The development of new high-throughput assays to read out regulatory functions other than enhancing gene expression (such as repression or insulation) will help as well.
Another missing piece in the story of human regulatory evolution is the vast sequence differences encoded within SVs, many of which are missing or incorrect in human and non-human primate reference genomes. Even within the SVs that have been accurately compared across primates, human non-coding evolution is relatively unexplored, due largely to the technical challenges of functionally testing regions with duplications and complex genomic architectures. SVs have immense potential to alter regulatory elements and their interactions with target genes. For example, inserting or deleting topologically associating domain (TAD) boundaries, which are often conserved across cell types and species, can associate regulatory elements with new genes (termed ‘enhancer hijacking’) or insulate them from their ancestral gene targets. It will be exciting to explore the evolution of human gene expression in terms of 3D genome organization and its effects on regulatory interactions.
이 글은 위 출처의 원문을 그 라이선스(CC BY 4.0)에 따라 옮겼습니다. 참고문헌 목록과 인용 번호, 그림 · 표 · 각주 · 수식 · 부록은 빼고 절별로 나누어 실었으며, 그 밖의 문장은 원문 그대로입니다. 원 저작자와 발행 기관은 이 화면을 만들거나 검토하지 않았고, 이곳의 해석과 낱말 풀이는 공식 번역이 아닙니다.
