NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Lodish H, Berk A, Zipursky SL, et al. Molecular Cell Biology. 4th edition. New York: W. H. Freeman; 2000.

  • By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.
Cover of Molecular Cell Biology

Molecular Cell Biology. 4th edition.

Show details

Section 14.2Cell-Type Specification in Animals

Each of the hundreds of different cell types found in multicellular organisms must be generated in the right number and in the appropriate region of the developing embryo and must be integrated into the framework of other cells to form discrete tissues. Specialized cells often have a distinctive morphology and express proteins devoted to the specific biochemical functions carried out by a particular cell type. The extensive cell specification that occurs during development of animals and plants depends on both quantitative and qualitative differences in gene expression, controlled largely at the level of transcription. An impressive array of molecular strategies, some analogous to those found in yeast cell-type specification, have evolved to carry out the complex developmental pathways that characterize multicellular organisms.

Cell biologists do not yet understand the complete set of regulatory molecules for any unique cell type in a multicellular organism that makes it different from other cells. In recent years, however, a family of related regulatory proteins have been shown to play analogous roles in the development of skeletal muscle cells (myogenesis), neural cells (neurogenesis), and perhaps other cell types. Mammalian skeletal muscle is a favorable system for investigating the role of transcription factors in controlling cell-type specification because its development can be studied both in the intact organism and in in vitro systems. In this section, we first describe the network of proteins that control specification of muscle cells and then consider how similar mechanisms operate during neurogenesis.

Embryonic Somites Give Rise to Myoblasts, the Precursors of Skeletal Muscle Cells

Mammalian skeletal myogenesis proceeds through three stages (Figure 14-8): determination of precursor muscle cells, called myoblasts; proliferation and in some cases migration of myoblasts; and their subsequent differentiation into mature muscle. In the first stage, myoblasts arise from blocks of mesodermal cells, called somites, found lateral to the neural tube in the embryo (Figure 14-9). Somites also give rise to tissues other than muscle including skeletal tissue and connective tissue in the skin. Specific signals from surrounding tissue including the neural tube and the lateral ectoderm play an important role in determining where myoblasts will form in the developing somite.

Figure 14-8. Schematic diagram of three stages in development of skeletal muscle in mammals.

Figure 14-8

Schematic diagram of three stages in development of skeletal muscle in mammals. Somites are collections of embryonic mesodermal cells, some of which become determined as myoblasts. Myoblasts, which are precursor skeletal muscle cells, are distinct from (more...)

Figure 14-9. Embryonic determination and migration of myoblasts in mammals.

Figure 14-9

Embryonic determination and migration of myoblasts in mammals. (a) Skeletal muscle is derived from embryonic structures called somites, which are blocks of mesodermal cells. (b) After formation of the neural tube, each somite forms a dermomyotome, which (more...)

Myoblasts are committed to become muscle but have not yet differentiated; hence, they are referred to as determined. For instance, myoblasts that will form muscles in the limb migrate from the lateral side of the myotome to the developing limb bud. Here the cells align, stop dividing, fuse to form a syncytium (a cell containing many nuclei but sharing a common cytoplasm), and differentiate into muscle. We refer to this multinucleate skeletal muscle cell as a myotube. Concomitant with cellular fusion there is a dramatic rise in the expression of genes necessary for muscle development and function. Other myoblasts from the more dorsal and medial regions of the myotome do not migrate and instead form cells of trunk muscles.

The specific extracellular signals that induce determination of each group of myoblasts are expressed only transiently. These signals trigger expression of numerous intracellular factors that can maintain the myogenic program in the absence of the inducing signals. We discuss the identification and functions of these myogenic proteins, and their interactions, in the next several sections.

Myogenic Genes Were First Identified in Studies with Cultured Fibroblasts

In vitro studies with the fibroblast cell line designated C3H 10T1/2 have played a central role in dissecting the transcription-control mechanisms regulating skeletal myogenesis. When these cells are incubated in the presence of 5-azacytidine, a cytidine derivative that cannot be methylated, they differentiate into myotubes. Upon entry into cells, 5-azacytidine is converted to 5-azadeoxycytidine triphosphate and then is incorporated into DNA in place of deoxycytidine. As noted in Chapter 10, methylated deoxycytidine residues commonly are present in transcriptionally inactive DNA regions. Thus replacement of cytidine residues with a derivative that cannot be methylated may permit activation of genes previously repressed by methylation. The high frequency at which treated C3H 10T1/2 cells are converted into myotubes suggested to early workers that reactivation of one or a small number of closely linked genes is sufficient to drive a program of muscle development.

To test this hypothesis, researchers isolated DNA from C3H 10T1/2 cells grown in the presence of 5-azacytidine, so-called azamyoblasts, and transfected it into untreated cells (Figure 14-10). The observation that 1 in 104 cells transfected with DNA isolated from azamyoblasts was converted into myotubes was consistent with the hypothesis that one or a small set of closely linked genes is responsible for converting fibroblasts into myotubes.

Figure 14-10. Experimental system for studying mammalian myogenesis.

Figure 14-10

Experimental system for studying mammalian myogenesis. A fibroblast cell line called C3H 10T1/2 can be converted into muscle cells by incubating them with 5-azacytidine. Under appropriate conditions, intermediate precursor cells, termed azamyoblasts, accumulate. (more...)

Subsequent studies led to the isolation and characterization of four genes that can convert C3H 10T1/2 cells into muscle. Figure 14-11 outlines the experimental protocol for identifying and assaying one of these genes, called myoD for myogenic determination gene D. Colonies of myoD-transfected cells were indistinguishable from C3H 10T1/2 cells treated with 5-azacytidine, and both types of cells exhibited myotube-like properties. The myoD cDNA also was found to convert a number of other cell lines into muscle. Based on these findings, the myoD gene was proposed to play a key role in muscle development. A similar approach has identified three other genes — the myogenin, myf5, and mrf4 genes — that also function in muscle development. As discussed in a later section, knockout experiments have demonstrated the importance of three of these genes in muscle development in the intact mouse.

Figure 14-11. Identification and assay of genes that drive myogenesis.

Figure 14-11

Identification and assay of genes that drive myogenesis. (a) Azamyoblast mRNAs were isolated from cell extracts on an oligo-dT column (see Figure 7-14). Incubation with reverse transcriptase and [32P]dNTPs yielded radiolabeled cDNAs. The cDNAs were mixed (more...)

Myogenic Proteins Are Transcription Factors Containing a Common bHLH Domain

The four myogenic proteins — MyoD, Myf5, myogenin, and Mrf4 — are all members of the basic helix-loop-helix (bHLH) family of DNA-binding transcription factors (see Figure 10-44). Near the center of these proteins is a DNA-binding basic region adjacent to the HLH domain, which mediates dimer formation. Flanking this central DNAbinding/dimerization region are two activation domains. We refer to the four myogenic bHLH proteins collectively as muscle-regulatory factors, or MRFs (Figure 14-12a).

Figure 14-12. Schematic diagrams of the general structures of two classes of transcription factors that participate in myogenesis.

Figure 14-12

Schematic diagrams of the general structures of two classes of transcription factors that participate in myogenesis. MRFs are produced only in muscle, whereas MEFs are expressed in several tissues in addition to developing muscle. The myogenic activity (more...)

bHLH proteins form homo- and heterodimers that bind to a 6-bp DNA site with the consensus sequence C-A-N-N-T-G. Referred to as the E box, this sequence is present in many different locations within the genome (i.e., on a purely random basis the E box will be found every 256 nucleotides). Thus some mechanism(s) must ensure that MRFs specifically regulate muscle-specific genes and not other genes containing E boxes in their transcription-control regions. We will examine how this myogenic specificity may be achieved using MyoD as an example.

The affinity of MyoD for DNA is tenfold greater when it binds as a heterodimer complexed with E2A, another bHLH protein, than when it binds as a homodimer. Indeed, in developing azamyoblasts, MyoD is found as a heterodimer complexed with E2A, and E2A, as well as MyoD, has been shown to be required for myogenesis in C3H 10T1/2cells. The DNA-binding domains of E2A and MyoD have similar but not identical amino acid sequences, and both proteins recognize the E-box sequence in DNA. However, since E2A is expressed in other tissues, the requirement for E2A is not sufficient to confer myogenic specificity.

MyoD stimulates transcription only when two or more molecules of MyoD are bound to multiple E boxes; this multiple binding occurs cooperatively. Although multiple copies of E-box sequences are found in most muscle-specific enhancers, they also are present in enhancers that promote expression of genes specifically expressed in other tissues. For instance, E2A is required for normal development of B cells (the white blood cells that produce antibodies), and it regulates B-cell specific genes through enhancers containing multiple E boxes. Thus neither the requirement for multiple E boxes nor the requirement for E2A is sufficient to confer myogenic specificity. Rather, as in yeast cell-type specification, the key to myogenic specificity lies in the combinatorial association of different transcription factors at different transcription-control sites.

MEFs Function in Concert with MRFs to Confer Myogenic Specificity

Some insight into how skeletal muscle cells are specified has come from in vitro mutagenesis studies in which variant E2A proteins were produced. Wild-type E2A protein cannot by itself drive C3H 10T1/2 cells to a myogenic fate, although it can bind to E boxes controlling muscle-specific genes. To identify which features of MyoD confer myogenic specificity, researchers produced altered E2A molecules in which specific amino acids present in MyoD were substituted into the equivalent positions in the E2A molecule. An E2A variant with three amino acid substitutions corresponding to residues present in MyoD was found to convert C3H 10T1/2 cells to myotubes. Two of these substitutions are in the central basic DNA-binding region of E2A, and one is just adjacent to this region. Although these substitutions allow E2A to drive myogenic conversion, they do not affect the DNA-binding specificity of E2A. This finding suggests that myogenic specificity is likely to reside in specific interactions between MyoD and other proteins. Recent studies indicate that specific amino acids in the bHLH domain of all the MRFs may confer myogenic specificity by allowing MRF-E2A complexes to bind specifically to another family of DNA-binding proteins called muscle enhancer – binding factors, or MEFs.

First identified in biochemical experiments as homo-dimeric proteins that bind to certain DNA sequences in a muscle-specific enhancer, MEFs later were shown to belong to the MADS family of transcription factors. In addition to containing an N-terminal MADS domain, these proteins contain a short stretch of amino acids just C-terminal to the MADS domain, called the MEF domain, and a C-terminal transcription-activation domain (Figure 14-12b). MEFs were considered excellent candidates for interaction with MRFs for two reasons: First, many muscle-specific genes contain both MEF –  and MRF – recognition sequences; second, although MEFs cannot induce myogenic conversion of C3H 10T1/2 cells, they enhance the ability of MRFs to do so. This enhancement requires physical interaction between a MEF and MRF-E2A heterodimer.

The interaction between these different transcription factors requires both the MADS and MEF domains in the MEF homodimer and the myogenic-specific amino acids in the bHLH domain of the MRF-E2A heterodimer. Crystallographic analysis of MRF-E2A bound to DNA indicates that these amino acids are buried within the major groove of DNA and are unable to directly contact MEFs. Hence, it seems likely that these amino acids confer a particular conformation to other regions of the MRF-E2A heterodimer that, in turn, specifically interact with MEFs. Surprisingly, although both classes of proteins can bind individually to specific DNA sequences, a single DNA site that recognizes one or the other protein is sufficient to act as a platform for assembly of a MRF-E2A-MEF complex. This finding suggests that different configurations of the DNA sites recognized by these factors may drive high levels of muscle-specific gene expression. Indeed, some muscle-specific genes contain bHLH-binding sites (i.e., E boxes); others contain MEF-binding sites; and yet others contain both types of protein-binding sites (Figure 14-13). In contrast to MRFs, which are expressed only in developing muscle, MEFs are expressed in other tissues including the developing central nervous system.

Figure 14-13. An MRF-E2A-MEF complex can assemble in the transcription-control region of muscle-specific genes containing an E box, MEF box, or both.

Figure 14-13

An MRF-E2A-MEF complex can assemble in the transcription-control region of muscle-specific genes containing an E box, MEF box, or both. The synergistic action of the MEF homodimer and MRF-E2A heterodimer, which directly interact, drives high-level expression (more...)

Myogenic Stages at Which MRFs and MEFs Function in Vivo Have Been Identified

Expression of any one of the four MRFs in C3H 10T1/2 cells can induce the cells to differentiate into muscle in vitro. The functions of these proteins in the intact animal during normal myogenesis have been studied in gene-knockout experiments. In these studies, mice were prepared with gene-targeted knockout mutations in the genes encoding MyoD, Myf5, or myogenin (see Figure 8-34). By analyzing the effects of knocking out these genes, developmental biologists could determine which genes are required for myogenesis and the stage at which they act.

Mice with either the myoD or myf5 gene knocked out have normal muscle, whereas those with the myogenin gene knocked out are missing the vast majority of skeletal muscle (Table 14-1). In mice that lack myogenin, myoblasts accumulate at sites normally occupied by skeletal muscle, indicating that myogenin is not required for formation of myoblasts but is required for their differentiation into myotubes. The simple, but erroneous, conclusion from these findings is that Myf5 and MyoD are not required for muscle development. However, since either protein can drive a myogenic program in cell culture, the loss of one gene may be compensated by the function of the other. Indeed, mice homozygous for mutations in both myf5 and myoD die shortly after birth and lack skeletal muscle. In contrast to the myogenin mutants, myoblasts do not accumulate in the myf5; myoD double mutants, suggesting that the Myf5 and MyoD proteins are required for the formation or survival of myoblasts. Overlapping functions such as these of MyoD and Myf5 are often referred to as redundant. Redundancy provides a more robust developmental program and may allow for more flexibility in the response of cells in different regions of the developing organism to extracellular signals regulating myogenesis.

Table 14-1. Effect of Knockout of Myogenic Genes in Mice.

Table 14-1

Effect of Knockout of Myogenic Genes in Mice.

The results of these gene-knockout experiments are consistent with the observation that azacytidine-treated C3H 10T1/2 cells express Myf5 and MyoD prior to fusion but express myogenin only as they fuse to form a syncytium, which then differentiates to form a myotube. As the model in Figure 14-14 illustrates, MyoD and Myf5 are thought to have similar but overlapping functions in selecting cells from developing somites to become myoblasts; that is, they are required for myoblast determination during normal myogenesis. Myogenin, then, is required for the differentiation of myoblasts into myotubes. The fourth MRF protein, Mrf4, is expressed later in development and may play a role in the maintenance of muscle cells.

Figure 14-14. Model of genetic control of mammalian skeletal muscle in vivo based on knockout experiments in mice and loss-of-function mutations in Drosophila.

Figure 14-14

Model of genetic control of mammalian skeletal muscle in vivo based on knockout experiments in mice and loss-of-function mutations in Drosophila. According to this model, MyoD and Myf5 serve a redundant function in myoblast determination, while myogenin (more...)

The role of E2A and MEFs in myogenesis have been assessed in more recent studies. Muscle development is normal in mice with a knockout mutation in the gene encoding E2A, although B-cell development is disrupted. Presumably, during muscle development, E2A-related genes may compensate for loss of E2A, much as myoD and myf5 can compensate for each other. To assess this redundancy, researchers will have to knock out the E2A-related genes and generate mice lacking both E2A and its related genes.

Because mice express multiple MEF proteins, scientists turned to Drosophila, which expresses a single MEF, to determine the function of MEFs in muscle development. In flies carrying loss-of-function mutations in the MEF gene, no differentiated muscle forms, although myoblasts appear to form normally. Hence, MEFs are required for differentiation, but not for determination (see Figure 14-14).

Multiple MRFs Exhibit Functional Diversity and Permit Flexibility in Regulating Development

The expression of four myogenic bHLH proteins (MRFs) in mice raises intriguing questions. Do these proteins have intrinsically different biochemical properties that correlate with distinctive roles in muscle development? That is, did functionally different MRFs evolve independently? Or have multiple MRFs evolved to facilitate the demands of gene expression in more complex organisms? That is, was duplication of an ancestral MRF gene and the subsequent evolution of divergent transcription-control elements more efficient than incorporation of different control elements into a single gene? Since many mouse genes that regulate development are found in multiple copies, understanding the role of the apparent duplication of the MRFs may provide generally applicable insights about developmental processes.

Scientists have begun to assess these possibilities using a variation of gene-knockout technology called knockin. In this technique, the coding sequences of one gene (e.g., myf5) are replaced by those of another (e.g., the myogenin gene). The experiments combining knockout and knockin technology depicted in Figure 14-15 demonstrate that myogenin and Myf5 are not functionally equivalent in vivo. Indeed, recent biochemical studies have shown that the chromatin-remodeling ability of Myf5 (and MyoD) is much greater than that of myogenin. As we discuss in more detail later, remodeling of chromatin is critical for normal development of most tissues.

Figure 14-15. Experimental demonstration that myogenin cannot substitute for Myf5 in vivo.

Figure 14-15

Experimental demonstration that myogenin cannot substitute for Myf5 in vivo. Expression of either Myf5 or myogenin in C3H 10T1/2 cells can drive myogenesis. To test whether these proteins are functionally equivalent in vivo, researchers inserted the myogenin-coding sequences (more...)

As noted earlier, mice with a homozygous knockout of the myogenin gene accumulate myoblasts and are not viable (see Table 14-1). By creating mice homozygous for the myogenin knockout and carrying one copy of the myogenin knockin at the myf5 locus (i.e., under control of the myf5 regulatory sequences), scientists could assess the importance of the myogenin-specific transcriptional regulation. The failure of this knockin to rescue the myogenin defect indicated that the unique expression pattern conferred by the myogenin regulatory sequences is also critical for normal development. In summary, these studies suggest that gene duplication led to evolution of genes encoding functionally diverse MRFs whose expression is regulated by different transcription-control elements.

Terminal Differentiation of Myoblasts Is under Positive and Negative Control

The determined, yet undifferentiated, myoblast can respond to extracellular signals in the developing embryo that control proliferation (hence, the number of cells that form) and cell migration (hence, the precise location of muscle). In contrast, the differentiated muscle cell, or myotube, cannot respond to such signals. Regulation of the transition from the determined to the differentiated state thus permits the precise spatial and temporal control of cellular differentiation that is necessary to ensure normal morphogenesis in complex multicellular organisms. The factors that regulate this critical step in various developmental pathways are still poorly understood. However, in vitro experiments have revealed several specific factors that promote or inhibit differentiation during myogenesis.

Inhibitory Proteins

Screens for genes related to myoD led to identification of a related protein that retains the dimerization helices but lacks the DNA-binding basic region and hence is unable to bind to E-box sequences in DNA. However, this protein interacts with MyoD and E2A, thereby inhibiting formation of MyoD-E2A heterodimers and hence their high-affinity binding to DNA. Accordingly, this protein is referred to as Id for inhibitor of DNA binding. Analysis of DNA from proliferating azamyoblasts, which express MyoD, E2A, and Id, has shown that the MyoD-binding (or E2A-binding) site in the promoter of the muscle-specific gene encoding creatine kinase is not occupied. This finding presumably reflects the formation of inactive MyoD-Id or E2A-Id complexes and indicates that Id can maintain cells in a determined state during proliferative growth. When these cells are induced to differentiate into muscle (for instance by the removal of serum-containing growth factors required for proliferative growth), the Id concentration falls. As a result, MyoD-E2A dimers can form and bind to the promoters of target genes driving differentiation of azamyoblasts into myotubes. We can see from these results that dimerization of transcription factors with different partners not only can modulate the specificity or affinity of their binding to specific DNA sites, but also may prevent their binding entirely.

Cell-Cycle Proteins

The onset of differentiation in many cell types is associated with arrest of the cell cycle, most commonly in G1, suggesting that cell-cycle proteins (e.g., cyclins and Cdks) may influence the transition from the determined to differentiated state. Researchers recently have found that certain inhibitors of cyclin-Cdk protein kinase activity can induce muscle differentiation in cell culture and that these inhibitors are markedly up-regulated in differentiating muscles in vivo. Conversely, differentiation of cultured myoblasts, under conditions in which they would normally differentiate, can be inhibited by transfecting the cells with DNA encoding cyclin D1 under the control of a constitutively active promoter. Expression of cyclin D1, which normally occurs only during G1, is up-regulated by mitogenic factors in many cell types (see Figure 13-29). The ability of cyclin D1 to prevent myoblast differentiation in vitro may mimic aspects of the in vivo signals that antagonize the differentiation pathway. The antagonism between negative and positive regulators of G1 progression is likely to play an important role in controlling myogenesis in vivo.

A Network of Cross-Regulatory Interactions Maintains the Myogenic Program

Precursor cells in different regions of the myotome give rise to different muscles: dorsal medial precursors to epaxial muscles, lateral precursors to hypaxial muscles, and ventrolateral precursors (after migrating) to limb muscles (see Figure 14-9). Each group of precursor cells shows a distinct pathway of myogenic gene activation induced by different signals from surrounding tissues. Once the myogenic program is activated in a region of the somite, an extensive array of cross-regulatory interactions acts to maintain it (Figure 14-16). These cross-regulatory interactions occur at two levels. First, myogenic factors, both MRFs and MEFs, positively regulate each other’s expression by binding to cis-acting regulatory sites. Second, MEFs and MRFs physically interact, thereby acting synergistically to promote expression of myogenic factors that drive differentiation. Thus, although the myogenic program is induced by extracellular signals transiently expressed in tissues surrounding the somite, a network of intracellular interactions maintains the myogenic program in the absence of these signals.

Figure 14-16. Maintenance of the myogenic program.

Figure 14-16

Maintenance of the myogenic program. Transient signals from the developing spinal cord and ectoderm induce a subset of cells in the developing somite to become myoblasts. Induction is marked by the expression of MRFs and MEFs. These proteins cross-regulate (more...)

Neurogenesis Requires Regulatory Proteins Analogous to bHLH Myogenic Proteins

Four bHLH proteins that are remarkably similar to the myogenic bHLH proteins control neurogenesis in Drosophila. These Drosophila proteins are encoded by an ≈100-kb stretch of genomic DNA, termed the achaete-scute complex (AS-C), containing four genes designated achaete (ac), scute (sc), lethal of scute (l’sc), and asense (a). Analysis of the effects of loss-of-function mutations indicate that the Achaete (Ac) and Scute (Sc) proteins participate in determination of neural precursor cells, while the Asense (As) protein is required for neural differentiation. These functions are analogous to the roles of MyoD and Myf5 in muscle determination and of myogenin in differentiation (Figure 14-17). Two other Drosophila proteins, designated Da and Emc, are analogous in structure and function to mammalian E2A and Id, respectively. For example, heterodimeric complexes of Da with Ac or Sc bind to DNA better than the homodimeric forms of Ac and Sc. Emc, like Id, lacks a DNA-binding basic domain; it binds to Ac and Sc proteins, thus inhibiting their association with Da and binding to DNA.

Figure 14-17. Comparison of genes that regulate Drosophila neurogenesis and mammalian myogenesis.

Figure 14-17

Comparison of genes that regulate Drosophila neurogenesis and mammalian myogenesis. bHLH transcription factors have analogous functions in precursor determination and subsequent differentiation into mature muscle cells and neurons. In both cases, the (more...)

A family of bHLH proteins related to the Drosophila Achaete and Scute proteins have been identified in vertebrates. One of these, called neurogenin, which has been isolated from the rat, mouse, and frog, appears to function in determination of neuronal precursor cells. In situ hybridization experiments have shown that neurogenin is expressed at an early stage in the developing nervous system and may induce expression of NeuroD, another bHLH protein that acts later (Figure 14-18a). Injection of large amounts of neurogenin mRNA into Xenopus embryos further demonstrated the ability of neurogenin to induce neurogenesis (Figure 14-18b). These studies suggest that the function of neurogenin is analogous to that of the Achaete and Scute in Drosophila; likewise, NeuroD and Asense may have analogous functions in vertebrates and Drosophila.

Figure 14-18. Experimental demonstration that neurogenin acts before NeuroD in vertebrate neurogenesis.

Figure 14-18

Experimental demonstration that neurogenin acts before NeuroD in vertebrate neurogenesis. (a) Neurogenin mRNA and neuroD mRNA were detected in the rat neural tube by in situ hybridization. Neurogenin mRNA is expressed in proliferating neuronal precursor (more...)

Gene knockout studies in mice, which have two neurogenin genes, have confirmed the essential role of neurogenin in vertebrate neurogenesis. In mice embryos that cannot express neurogenin-1, the trigeminal ganglion in the head region does not develop. However, other regions of the nervous system develop normally in neurogenin-1 knockouts, suggesting that neurogenin-2 or other bHLH proteins regulate neurogenesis in these regions. In the region of the nervous system affected by the loss of neurogenin-1, development is arrested before expression of NeuroD begins.

Progressive Restriction of Neural Potential Requires Inhibitory HLH Proteins and Local Cell-Cell Interactions

The regulatory mechanisms responsible for restriction of particular developmental pathways to specific cells within an embryo are very complex and not thoroughly understood for any system. For instance, specific cells within somites are selected to become myoblasts, while other cells are destined to become nonmuscle tissues (see Figure 14-9). The bestunderstood example of such developmental restriction occurs during formation of sensory bristles in Drosophila. In this case, the proteins encoded by the proneural genes achaete and scute must be expressed and active in cells selected to become neural precursor cells, but not in surrounding cells.

The sensory bristles located on the second thoracic segment of the adult fly arise from a monolayer of cells (a columnar epithelium) called the wing imaginal disc, from which the epidermis of this segment and the associated wing also are derived. Each bristle is part of a sensory organ that contains four cells: a bristle that protrudes from the epidermis; the socket into which it is inserted; a neuron that transmits the sensory information; and, finally, a cell associated with the neuron referred to as a support cell. These cells are derived from a single cell by two sequential divisions. This “grandmother cell” is referred to as a sensory organ precursor (SOP), or more generally as a neural precursor cell. The pattern of SOPs in the developing imaginal disc presages the pattern of bristles in the adult, which is highly reproducible. These cells do not migrate from another location, but rather arise in distinct positions within the columnar epithelium of the imaginal disc. Each SOP emerges from a cluster of cells, the proneural cluster, that express the Achaete and Scute proteins. The pattern of proneural clusters (i.e., of Ac and Sc expression) is determined by earlier-acting patterning genes whose encoded proteins function to divide the epithelium into developmental domains. During normal development, only one cell in each proneural cluster is selected to become an SOP; the remaining cells develop into epidermal structures. Restriction of the neural program to one cell results from down-regulation of the activity and expression of the proneural proteins Achaete and Scute.

As noted earlier, Emc inhibits the binding of Achaete and Scute to DNA and hence their ability to determine neural precursor cells (see Figure 14-17). Loss-of-function mutations in emc lead to formation of multiple sensory bristles from a single proneural cluster, whereas gain-of-function mutations suppress SOP formation. In wild-type embryos, expression of Emc is lower in the region of each proneural cluster from which a SOP will arise than in the regions giving rise to epidermal structures. Like achaete and scute, emc is under complex regulation so that Emc is expressed in a specific pattern within the developing epithelium. The resulting variation in the relative level of proteins promoting neurogenesis (e.g., Achaete and/or Scute) and the level of proteins inhibiting it (e.g., Emc) in the cells of the wing imaginal disc limits SOP-forming potential to a small group of neighboring cells within each proneural cluster (Figure 14-19).

Figure 14-19. Formation of sensory organ precursors (SOPs) in the wing imaginal disc of Drosophila.

Figure 14-19

Formation of sensory organ precursors (SOPs) in the wing imaginal disc of Drosophila. A set of extracellular signaling molecules and transcription factors, encoded by so-called patterning genes, control the precise spatiotemporal pattern of expression (more...)

Short-range cell-cell interactions further restrict SOP formation by inhibiting the expression of Achaete and/or Scute in all but one cell of a proneural cluster. These local interactions are mediated by two cell-surface proteins: Notch, a cell-surface receptor, and Delta, its specific ligand (Chapter 23). Loss-of-function mutations in either the Notch or Delta locus result in the formation of multiple SOPs from a proneural cluster and the appearance of multiple bristles arising from a single proneural cluster, indicating that these genes act to inhibit SOP formation (Figure 14-20). Local asymmetry in the expression of Notch and Delta, reinforced by variations in the levels of Achaete, Scute, and Emc, permits one cell, and only one cell, in a proneural cluster to retain its neural potential and become an SOP (Chapter 23; see Figure 23-28). The selected SOP then begins to express Asense and other differentiation proteins, which determine the type of neuron that develops, a process termed neuronal specification. As production of Asense protein increases in an SOP, synthesis of Achaete and Scute proteins decreases.

Figure 14-20. Cell-cell interactions that down-regulate proneural genes are critical in determination of a single SOP within a proneural cluster.

Figure 14-20

Cell-cell interactions that down-regulate proneural genes are critical in determination of a single SOP within a proneural cluster. Notch, a cell-surface receptor, and Delta, another cell-surface protein that binds to Notch, are initially expressed in (more...)

bHLH Regulatory Circuitry May Operate to Specify Other Cell Types

The important role of bHLH proteins in myogenesis and neurogenesis is supported by discovery of similar highly conserved regulatory proteins in C. elegans. Moreover, considerable evidence indicates that a bHLH protein, called SCL, participates in determination of hematopoietic stem cells, which differentiate to generate the many different types of blood cells. SCL is expressed in the ventral mesoderm of the developing embryo, in a region of the mesoderm giving rise to hematopoietic stem cells; like MyoD and Myf5, SCL forms a complex with the more generally expressed E2A protein.

The specification of cell type may be an ancient function of bHLH proteins. In the cnidarian Hydra vulgaris, which diverged from arthropods some 600 million years ago, a bHLH protein is specifically expressed in the nematocyte, one of some 20 different cell types found in this organism. Nevertheless, it is unlikely that all cell types will be controlled by this regulatory circuitry. Future research of diverse cell types will most likely uncover additional strategies of cell specification using different networks of transcription factors.


  •  In some systems, transient extracellular signals induce a cell-specification program, and an intracellular network of regulatory proteins maintains it.
  •  Skeletal muscle cells arise from a subclass of cells in the developing somite that are induced to express myogenic bHLH proteins, or MRFs. MyoD and Myf5 are required for commitment of mesodermal cells to myoblasts, and myogenin is required for myoblasts to differentiate into myotubes (see Figure 14-14).
  •  MEFs and MRFs bind to each other and act synergistically to control transcription of muscle-specific genes.
  •  Since myoblasts continue to proliferate and, in some cases, migrate to different regions of the developing embryo, specific mechanisms must maintain the determined state and prevent differentiation until an appropriate time. These mechanisms include inhibitory proteins (e.g., Id) that prevent the formation of bHLH dimers and proteins that promote cell-cycle progression.
  •  Neurogenesis in flies is controlled by a network of bHLH proteins analogous to those controlling skeletal myogenesis (see Figure 14-17). A similar network probably controls vertebrate neurogenesis.
  •  Neuronal precursor cells (SOPs) in Drosophila arise from equipotent groups of cells, called proneural clusters, which all express Achaete and/or Scute. The ability of these cells to give rise to neuronal precursors is progressively restricted (see Figure 14-19).
  •  Emc inhibits expression of Achaete/Scute in many cells in a proneural cluster, leaving a small number of cells competent to form SOPs. Interactions between these cells mediated by two transmembrane proteins restricts neuronal specification to a single cell (see Figure 14-20).
Image ch7f14
Image permission
Image ch8f34
Image ch13f29
Image ch23f28

By agreement with the publisher, this book is accessible by the search feature, but cannot be browsed.

Copyright © 2000, W. H. Freeman and Company.
Bookshelf ID: NBK21618