Harvard University COVID-19 updates

Department News



by the astonishing variety of forms, shapes, and colors that have evolved in nature. This variation has allowed organisms to adapt to their local environments, for example, to evade predators or attract suitable mates. Biologists have long studied the ultimate causes underlying this diversity (i.e., the why), but the mechanisms of its origin (i.e., the how) have been less well studied. The challenge is to study mechanism in non-model species. Non-model organisms have tremendous morphological diversity that has evolved naturally, but are less tractable compared to model organisms for which a wealth of genetic and molecular tools has been developed.

Our study, published in Nature  (PDF), shows how recent advances in genomic and molecular approaches offer exciting opportunities to understand the mechanistic basis of fitness-related traits. By taking advantage of the naturally evolved color pattern of the African striped mouse (Rhabdomys pumilio), we investigated the mechanisms responsible for forming periodic stripes, a common pattern in mammals. Striped mice are native to southern Africa, where they are highly social, diurnal rodents whose dorsal coat has four dark and two light longitudinal stripes arranged in a dark-light-dark pattern (Figure 1). Striped mice have three different types of hair: light (no pigment), dark (black pigment, eumelanin), and banded (eumelanin and yellow pigment, pheomelanin). We quantified the proportion of each hair type in different regions along the dorsoventral axis and showed that their proportions differed: the middle stripe and flank had similar proportions of all three hair types, but the light stripe contained mostly light hair and the dark stripe contained mostly dark hair.

Figure 1 The African striped mouse Rhabdomys pumilio (Kagalagadi, Botswana). Credit: J. F. Broekhuis

We next asked what gives rise to the differences in the spatial distribution of differentially pigmented hair? To understand how and when these differences in hair pigmentation arise, we analyzed striped mice from different stages of development and found that the stripe pattern is established during late embryogenesis. We then focused on these stages and asked: are the light stripes caused by a lack of melanocytes? We first examined the distribution of melanocytes, the only cell type in mammals capable of producing pigment. Although melanocytes were present in both light and dark stripes and at similar numbers, we noticed that those in light stripes failed to fully differentiate, which prevented them from producing pigment. As a first step to understanding why these melanocytes fail to differentiate, using RNA-sequencing, we compared the relative abundance of transcripts during the formation of light and dark stripes. A clear outlier emerged: Alx3, a homeodomain transcription factor was present at higher levels in light than dark stripes. At first, this result was puzzling because Alx3 was not previously implicated in regulating pigment. However, prior to the appearance of stripes, Alx3 was already present in the region of dorsal skin where light stripes would form.

The striking correlation between early Alx3 expression and later appearance of light stripes led us to hypothesize that this gene may be influencing melanocyte differentiation and thus the production of pigment. To test this directly, we used a technique developed by Connie Cepko and colleagues at Harvard Medical School. This involves delivering a lentivirus to developing laboratory mouse embryos in utero, using ultrasound-guided injections. First, we generated a virus containing Alx3 and injected it into the amniotic cavity of early stage embryos, prior to neural tube closure, which allowed us to target melanocyte progenitor cells. A few days later, when we analyzed the samples, we found that, indeed, those melanocytes infected with the Alx3-carrying virus were not able to terminally differentiate and produce pigment. Thus, this experiment recapitulated the process occurring during the formation of the light stripe, suggesting Alx3 alone can prevent melanocyte maturation. Second, to gain a more detailed understanding of the mechanism by which Alx3 alters melanocyte development, we used a combination of bioinformatic and biochemical approaches to show that Alx3 was directly repressing one of the key regulators of melanocyte differentiation – a transcription factor called Mitf  – by binding to its promoter. Thus, using a combination of in vivo and in vitro approaches, we found that spatial-specific expression of Alx3, via repression of Mitf, prevents the melanocyte maturation leading to the production of unpigmented hairs of dorsal light stripes.

To determine whether this mechanism might be ubiquitous, we analyzed skin samples from chipmunks, which last shared a common ancestor with African striped mice over 70mya and independently evolved a similar dorsal stripe pattern. Surprisingly, we found that Alx3 also is expressed at high levels in dorsal stripes consisting of light hair, showing that Alx3 is a patterning mechanism that has repeatedly evolved in different rodent groups.

Using African striped mice, we have identified a new mechanism to produce pigment patterning (Alx3 modulates color via inhibition of Mitf expression without loss of melanocytes). Thus, we advance our understanding of the formation and evolution of the remarkable array of mammalian coat color patterns found in nature, and gain insights into the molecular and cellular mechanisms by which phenotypic variation originates.

Our results lead to a number of additional questions that we are pursuing. For example, when comparing striped rodents, it is clear that stripes vary greatly in position (some species have stripes that are located more laterally than others), in number (some species have a single stripe along the midline and others have multiple stripes covering the dorsum), and in overall complexity (some species have both periodic stripes and spots). We are excited to investigate the mechanisms that regulate the spatial expression of Alx3. Ongoing work may lead to further insights into how mammals got their stripes.

The work was an international collaboration that included researchers from HudsonAlpha Institute for Biotechnology and Stanford University; Instituto de Investigaciones Biomédicas Alberto Sols (CSIC/UAM) and Ciber de Diabetes y Enfermedades Metabólicas Asociadas (Ciberdem); Collège de France; UNISTRA and University of the Witwatersrand.

Nature or download PDF

The Atlantic

The Scientist

Science News



Nature News & Views


Ricardo Mallarino (l) and Hopi Hoekstra

Ricardo Mallarino (l) and Hopi Hoekstra