RESEARCH NEWS & VIEWS evolutionary progression, but are alternative states that could, in theory, evolve back and forth. At the time, however, such transitions were unknown, and it seemed that each reptile species had one type of system or the other, like two sides of a coin. This latter perception is now also changing, and it seems that we are seeing the coin as it flips. The lizards studied by Holleley et al. (Fig. 1) have visibly recognizable sex chromo somes with female heterogamety — females have a Z and a W chromosome, and males have two Zs. However, the authors find that nearly 20% of ZZ individuals sampled in the wild are female instead of male. Incubation of eggs in the laboratory revealed that the ZZ offspring develop as male at low temperatures but that an increasing proportion develop as female as the incubation temperature increases. So ZZ females in the wild probably come from warm nests. These observations cement previous specu lation about how sex chromosomes and envi ronmental sex determination may coexist and how the transition between them may occur. At first glance, two problems are created by a system that combines sex chromosomes and environmental sex determination. The first is that arbitrary environmental determina tion of sex would lead to ZW males and ZW females, which when mated would sometimes lead to WW offspring. If the W is a degenerate sex chromosome, meaning it has lost many of its functional genes, WW offspring would be inviable or sterile. This is solved in P. vitticeps by the simple fact that ZW is always female — only the ZZ genotype becomes either sex — so there is no possibility of a ZW–ZW mating and thus no WW genotype. The second problem is that the consistent development of some ZZ individuals into females creates an excess of females in the population. This dilemma is solved through sex-ratio selection, which automatically adjusts the frequency of the W chromosome to progressively lower levels as more ZZ females are produced. Provided that the envi ronment is neither too warm nor too cool, the equilibrium population may sit indefinitely at a point that includes some ZW females and some ZZ females. But there is a continuum of equilibria spanning from pure chromosomal sex determination to pure environmental sex determination as the average nest temperature increases, and it is a steep transition. Extended Data Figure 4 of Holleley and colleagues’ paper1 shows that the wild population sits on a virtual cliff of the changeover between the two mechanisms. The study augments this picture with sev eral other observations. First, the tempera tures causing ZZ lizards to become female are slightly lower for offspring of ZZ moth ers than for offspring of ZW mothers. This result suggests underlying quantitative varia tion in the propensity for environmental sex
determination, a satisfying confirmation of theory: those individuals most genetically dis posed to develop as ZZ females have offspring that are also genetically disposed to become ZZ female. Second, ZZ females have markedly higher fecundity than ZW females. This result was not expected, and although it is not con tradictory to theory, it raises the question of why. The answer may shed light on the selec tive advantage of temperature-dependent sex determination. Holleley and colleagues’ findings will no doubt inspire parallel work on other species, especially in efforts to understand the transi tions between sex-determining mechanisms and to explore the ecological and evolution ary consequences of the different mechanisms. The ability to assess the fitness of ZZ and ZW females raised at the same temperature will enable comparisons to be made that are cru cial to understanding the relative advantages of the two systems and the possible costs of sex-chromosome degeneration. Broader geographic and longitudinal com parisons for these lizards will give insight into the ramifications of climate change on this temperature-dependent reproductive mode. However, the established equilibrium between genetically and environmentally determined sex in these lizards should respond quickly to climate change, because an overproduction of ZZ females in warm years would lead to a com pensatory reduction in the frequency of ZW
females in the next generation and beyond. The findings in this one system should dove tail with recent revelations of frequent changes in lizard heterogametic determination, a dis covery made possible by easy genome sequenc ing of less-studied species6. The accumulating information about the molecular bases of reptile sex determination7 will add greatly to this understanding, and may reveal interesting constraints imposed on the transitions8. The emerging picture is that reptilian sex determi nation is more flexible on an evolutionary scale than could ever have been imagined. ■ James J. Bull is at the Institute for Cellular and Molecular Biology, the Center for Computational Biology and Bioinformatics and the Department of Integrative Biology, University of Texas at Austin, Austin, Texas 78712, USA. e-mail:
[email protected] 1. Holleley, C. E. et al. Nature 523, 79–82 (2015). 2. Radder, R. S., Quinn, A. E., Georges, A., Sarre, S. D. & Shine, R. Biol. Lett. 4, 176–178 (2008). 3. Quinn, A. E. et al. Mol. Genet. Genomics 281, 665–672 (2009). 4. Ohno, S. Sex Chromosomes and Sex-linked Genes (Springer, 1967). 5. Janzen, F. J. & Paukstis, G. L. Q. Rev. Biol. 66, 149–179 (1991). 6. Gamble, T. et al. Mol. Biol. Evol. 32, 1296–1309 (2015). 7. Janes, D. E. et al. Biol. Lett. 10, 20140809 (2014). 8. Schwanz, L. E., Ezaz, T., Gruber, B. & Georges, A. J. Evol. Biol. 26, 2544–2557 (2013).
N EUR O BI O LOGY
Inversion in the worm Combinations of spatially and temporally restricted transcription factors are shown to coordinate movement in nematode worms by controlling the formation of synaptic connections to and from motor neurons. See Letter p.83 VILAIWAN M. FERNANDES & CLAUDE DESPLAN
N
ervous systems are staggeringly complex. To generate appropriate behavioural outputs, the countless synaptic connections that neurons form with other cells must be precisely regulated to ensure that they are arranged in the right cir cuits at the right time. On page 83 of this issue, Howell et al.1 show that, in the simple neuronal circuits of the nematode worm Caenorhabditis elegans, such spatio-temporal precision is achieved through transcription factors that function together to restrict the expression of a newly identified synaptic organizer protein called OIG-1. Nematode movement involves coordinated muscle contractions that are regulated by com plex interactions between motor neurons on
4 4 | N AT U R E | V O L 5 2 3 | 2 J U LY 2 0 1 5
© 2015 Macmillan Publishers Limited. All rights reserved
the worm’s dorsal (upper) and ventral (lower) sides. Two classes of inhibitory D-type motor neuron2 in particular help to control this pro cess. Dorsal D (DD) neurons make synaptic connections to muscles on the dorsal side of the worm, and are themselves innervated by excitatory motor neurons called cholinergic neurons from the animal’s ventral side. By contrast, ventral D (VD) neurons innervate ventral muscles and receive synaptic inputs from dorsal cholinergic neurons. Activa tion of ventral cholinergic neurons therefore contracts ventral muscles and activates DD neurons, inhibiting dorsal-muscle contrac tion, whereas activation of dorsal cholinergic neurons leads to contraction of dorsal, but not ventral, muscles. DD neurons form during embryonic devel opment, but VD neurons arise after a larva’s exoskeleton has moulted for the first time2.
NEWS & VIEWS RESEARCH a Early development
b
Muscle
After synaptic inversion
Dorsal Excitatory
UNC-30 and LIN-14
OIG-1
UNC-30
OIG-1
UNC-30 and UNC-55
OIG-1
Inhibitory VD
Inhibitory DD Ventral
Figure 1 | Transcriptional regulation of synaptic specificity. a, During early development of nematodes, excitatory motor neurons (blue) on the dorsal side of the embryo make synaptic connections with, and so excite, dorsal muscle and inhibitory dorsal D-type (DD) motor neurons (red). DD neurons innervate, and thus inhibit contraction of, ventral muscle. Howell et al.1 report that the synaptic inputs and outputs of early DD neurons are controlled by LIN-14. This transcription factor, together with the ubiquitously expressed transcription factor UNC-30, promotes expression of the protein OIG-1,
In fact, before VD neurons develop, early DD neurons innervate ventral muscles and receive synaptic inputs from motor neurons on the dorsal side of the animal2 (Fig. 1a). They later undergo a synaptic inversion when VD neu rons arise (Fig. 1b). Howell et al. investigated the regulatory logic behind this synaptic rewir ing using worms that harbour mutations in the gene unc-30, which encodes an evolutionarily conserved transcription factor, UNC-30. This protein is expressed in all D-type motor neu rons throughout development, and is regarded as a ‘terminal-identity selector’ — its expres sion ultimately defines whether neurons will become D-type motor neurons3–5. In addition to the uncoordinated locomo tion for which the gene is named, the authors found that unc-30 mutant worms had defects in synaptic specificity. Although the cell bod ies (the nucleus-containing regions) of D-type motor neurons are appropriately positioned in unc-30 mutants, the early DD and VD neurons fail to innervate ventral muscles and instead make abnormal synaptic connections to dorsal muscles. Moreover, VD neurons do not receive synaptic inputs from dorsal cholinergic motor neurons. How does UNC-30 regulate synapse forma tion in specific spatio-temporal patterns, given its ubiquitous expression in D-type neurons throughout development? Previous studies6–9 have shown that mutations in two other genes encoding transcription factors, lin-14 and unc-55, recapitulate different aspects of the defects found in unc-30 mutants. Expression of LIN-14 is temporally restricted to early devel opment, and mutation of lin-14 leads to the abnormal formation of dorsal synapses from early DD neurons8,9. By contrast, UNC-55 is restricted to VD neurons, where it regulates synaptic specificity to ventral muscles6,7. Howell et al. reasoned that UNC-30 might
which prevents the formation of synaptic outputs from DD neurons to dorsal muscle cells. b, After the larva’s first moult, LIN-14 is no longer expressed and DD neurons undergo a synaptic inversion — they become innervated by ventral excitatory neurons and themselves innervate dorsal muscle. Ventral D-type (VD) motor neurons (purple) innervate ventral muscle cells and express the transcription factor UNC-55, which, together with UNC-30, promotes OIG-1 expression and prevents VD neurons from forming inhibitory connections to dorsal muscle.
cooperate with LIN-14 and UNC-55 to pro mote the expression of a molecule that blocks dorsal synaptic output. This molecule would be expected to be expressed under the control of UNC-30 acting with LIN-14 in early DD neurons, and under the control of UNC-30 acting with UNC-55 in VD neurons, inhibit ing the formation of synaptic outputs on the dorsal side of the animal. But under this simple model, ventral synaptic outputs would remain normal in unc-30 mutants. Because this is not the case, additional layers of regulation must be involved. OIG-1, a member of the immunoglobulin superfamily, whose members mediate inter actions between cells, is a putative target of UNC-30, LIN-14 and UNC-55 (ref. 10). Howell and colleagues provide evidence that oig-1 is expressed in early DD neurons and is later restricted to VD neurons. Furthermore, its expression is altered by perturbations to lin-14 and unc-55 expression. The authors report that loss of oig-1 leads to the formation of dorsal synaptic outputs similar to those seen in unc-30, lin-14 and unc-55 mutants. More over, synaptic innervation of early DD and VD neurons by cholinergic neurons on the dorsal side of the animal is disrupted in oig-1 mutants. This suggests that OIG-1 coordinates both the inputs to and outputs of D-type motor neurons. Howell et al. found that OIG-1 is located along the ventral side of early DD and VD neurons. Given that the protein can organ ize dorsal synaptic inputs and outputs, this observation suggests that it acts indirectly. The authors also show that forced expression of oig-1 could not block synaptic rewiring of late DD motor neurons to ventral muscles, indicating that other factors must cooperate with OIG-1 to regulate synaptic specificity. Identifying these cofactors, which, like oig-1,
must be differentially expressed in early DD and VD neurons, will be of great interest, as will determining the molecular mechanisms by which they act with OIG-1 to regulate syn aptic inputs and outputs through pre- and postsynaptic partner molecules. Achieving appropriate synaptic specificity involves many developmental steps that act together to ensure that neurons assume the correct identity. Determinants of neuronal identity, which are regulated by terminalidentity selectors, include the production of particular neurotransmitter molecules, the guidance of nerve fibres in specific direc tions and the appropriate growth of branched projections called dendrites. Howell and col leagues’ work demonstrates that the formation and targeting of synapses are also traits that give neurons a particular identity, and that synapses, too, can be regulated by terminalidentity selectors such as UNC-30. ■ Vilaiwan M. Fernandes and Claude Desplan are in the Department of Biology, New York University, New York, New York 10003, USA. e-mail:
[email protected] 1. Howell, K., White, J. G. & Hobert, O. Nature 523, 83–87 (2015). 2. White, J. G., Albertson, D. G. & Anness, M. A. R. Nature 271, 764–766 (1978). 3. Eastman, C., Horvitz, H. R. & Jin, Y. J. Neurosci. 19, 6225–6234 (1999). 4. Jin, Y., Hoskins, R. & Horvitz, H. R. Nature 372, 780–783 (1994). 5. Cinar, H., Keles, S. & Jin, Y. Curr. Biol. 15, 340–346 (2005). 6. Walthall, W. W. & Plunkett, J. A. J. Neurosci. 15, 1035–1043 (1995). 7. Hedgecock, E. M., Culotti, J. G., Hall, D. H. & Stern, B. D. Development 100, 365–382 (1987). 8. Ruvkun, G. & Giusto, J. Nature 338, 313–319 (1989). 9. Hallam, S. J. & Jin, Y. Nature 395, 78–82 (1998). 10. Aurelio, O., Hall, D. H. & Hobert, O. Science 295, 686–690 (2002). 2 J U LY 2 0 1 5 | V O L 5 2 3 | N AT U R E | 4 5
© 2015 Macmillan Publishers Limited. All rights reserved
