SEARCH

SEARCH BY CITATION

Keywords:

  • neurogenesis;
  • differentiation;
  • receptor tyrosine kinase (RTK);
  • dorsal root ganglion;
  • neural epidermal growth-factor like like 2 (Nell2);
  • SSH-PCR

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The dorsal root ganglia (DRG) derive from a population of migrating neural crest cells that coalesce laterally to the neural tube. As the DRG matures, discrete cell types emerge from a pool of differentiating progenitor cells. To identify genes that regulate sensory genesis and differentiation, we have designed screens to identify members from families of known regulatory molecules such as receptor tyrosine kinases, and generated full-length and subtractive cDNA libraries between immature and mature DRG for identifying novel genes not previously implicated in DRG development. Several genes were identified in these analyses that belong to important regulatory gene families. Quantitative PCR confirmed differential expression of candidate cDNAs identified from the subtraction/differential screening. In situ hybridization further validated dynamic expression of several cDNAs identified in our screens. Our results demonstrate the utility of combining specific and general screening approaches for isolating key regulatory genes involved in the genesis and differentiation of discrete cell types and tissues within the classic embryonic chick model system. Developmental Dynamics 229:618–629, 2004. © 2004 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

A central question in biology is how a multicellular organism, consisting of an extremely diverse population of cell types, can arise from a single cell. Within the nervous system, this question is of particular interest when considering how the diverse cell types that comprise the peripheral nervous system (PNS) derive from a common pool of progenitor cells, the neural crest (Weston, 1970; LeDouarin, 1982; Knecht and Bronner-Fraser, 2002). With the advent of molecular strategies for conducting comprehensive analyses of gene expression, one can readily identify differential patterns of gene expression in a particular tissue of interest that can orchestrate the differentiation of discrete cell types. Application of such technologies expedites the search for key regulatory molecules, which has proven extremely powerful for identifying genes that play pivotal roles during neuronal development and repair (Su et al., 2002; Griffin et al., 2003; Tietjen et al., 2003). Additionally, combining differential screening strategies with recent gain-of-function and loss-of-function approaches returns the classic embryologic chick model system to the forefront of developmental biology (Swartz et al., 2001; Pekarik et al., 2003).

An excellent system within the PNS in which to address this question is the dorsal root ganglion (DRG), one of the major derivatives of the neural crest (Teillet et al., 1987; Lallier and Bronner-Fraser, 1988). Much is known about the biology of mature DRG: once sensory neurons differentiate, they innervate discrete central and peripheral targets, become dependent on particular neurotrophins for survival and undergo an extensive period of target-regulated programmed cell death resulting in the generation of approximately 20 distinct subclasses of sensory neurons and two types of glial cells (Scott, 1992; Lindsay, 1996; Snider and Silos-Santiago, 1996). In contrast, immature DRG contain a less well-defined population of cells consisting primarily of nascent, undifferentiated neurons and importantly, mitotically active progenitor cells that can generate neurons and/or glial cells (Pannese, 1974; Goldstein et al., 1995; Wakamatsu et al., 2000; Rifkin et al., 2000). Thus an elucidation of the cellular and molecular interactions that occur after neural crest migration yet before the completion of target innervation, is required because it is during this time period when all of neurogenesis and the onset of differentiation of discrete classes of sensory neurons occurs (Rifkin et al., 2000; Oakley and Karpinski, 2002; Guan and Condic, 2003). Several key regulatory molecules have been identified in sensory neural development, including neurotrophic factors and their receptors, transcription factors such as Neurogenin 1 and Neurogenin 2, ETS family members, and Notch-Delta and β-catenin signaling pathways (Ma et al., 1999; Lin et al., 1998; Arber et al., 2000; Wakamatsu et al., 2000; Patapoutian and Reichardt, 2001; Hari et al., 2002). Though these molecules and pathways play important roles during DRG development, they alone cannot fully account for the developmental processes underlying the genesis and differentiation of discrete cell types.

Thus the goal of this study was to identify new molecules involved in sensory neurogenesis and differentiation in the nascent DRG by searching for genes with a highly restricted developmental expression pattern. The identification of differentially expressed genes during development begins by either comparing the ensemble of transcripts or the ensemble of proteins from a particular immature tissue to those from the mature tissue. Previous DRG transcriptome analyses have been conducted (Akopian and Wood, 1995; Friedel et al., 1997; Dong et al., 2001; Costigan et al., 2002; Wang et al., 2003; Shin et al., 2003); however, none of these screens were designed to specifically identify molecules regulating early events in the immature DRG, i.e., during sensory neurogenesis and differentiation.

To this end, our strategies were designed to identify genes that function specifically during the peak period of sensory neurogenesis and differentiation within immature DRG. One strategy focused on identifying members of the receptor tyrosine kinase (RTK) gene family expressed in immature DRG. RTKs are an important class of proteins known to regulate cellular proliferation and differentiation in a myriad of cell and tissue types (Schlessinger, 2000). The second strategy used Suppressive Subtractive Hybridization (SSH)-PCR to directly compare gene expression in the immature chick DRG with that of the more mature DRG, rather than looking specifically for known gene families (Diatchenko et al., 1996). The genes identified in our screens were not found in previous DRG transcriptome analyses. Characterization of one of these genes, NELL2, not only demonstrated a pivotal role for NELL2 during DRG development (Nelson et al., 2002; Nelson et al., submitted), but also the feasibility and strategy for determining the function of genes identified in our screens. Our study has revealed several intriguing new molecules whose function during DRG development we are now in the process of determining.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

DRG derive from migrating neural crest cells that coalesce laterally to the neural tube beginning at ∼ Embryonic Day 2.75–3 in the chick (E2.75–E3; Teillet et al., 1987; Lallier and Bronner-Fraser, 1988). Neurogenesis in the nascent DRG then ensues, peaking at ∼E4.5–E5, which is followed by target innervation and programmed cell death of postmitotic neurons between ∼E5 and E12, peaking between E7 and E9 (Carr and Simpson, 1978). At E4.5, DRG are immature with ∼30% of the cells being mitotically active progenitor cells, and the majority of the remaining cells being nascent, postmitotic neurons (Goldstein et al., 1995; Rifkin et al., 2000; Wakamatsu et al., 2000). By E8.5 DRG neurogenesis is complete and the majority of neurons express their mature phenotypes (e.g., Oakley et al., 1997). Gliogenesis becomes prevalent after E6 (Carr and Simpson, 1978). Thus we chose E4.5 and E8.5 DRG as the ages to compare.

Construction of E4.5 DRG Full-Length cDNA Phage Library

We first generated a full-length cDNA phage library from E4.5 DRG that we could use to retrieve full-length clones resulting from our screens. The titer of the library was determined to be ∼1.0 × 108 plaque forming units (pfu)/ml. To determine the representational fidelity/complexity of the library, we screened and confirmed the expression of several genes known to be expressed in the DRG at this time including trkC, α-tubulin, and Brn3a (Table 1). For some of these products, cDNAs were up to 4 kb in length, demonstrating the quality of cDNAs in the library. We also screened it for expression of candidate genes of interest, in particular, a novel protocadherin (in collaboration with Roger Bradley) which we found was in fact present (data not shown). Protocadherins have been demonstrated to play important roles during development, are expressed in the DRG, and are part of quite a large family of cell adhesion molecules (Bradley et al., 1998; Heggem and Bradley, 2003). Identification of novel family members expressed during the peak of sensory neurogenesis and differentiation may be quite significant.

Table 1. Summary of Genes Identified in DRG Libraries and Screensa
Genes identified in full-length E4.5 DRG cDNA phage libraryGenBankMethod
Alpha-TubulinV00388Probing
TrkCS74248Probing
Bm3aCA786304Probing
Novel Protocadherin family memberIn progressPCR/Probing
Receptor tyrosine kinase RT-PCR screen of E4.5 DRG 
Platelet-Derived Growth Factor Receptor Alpha, PDGFACD216500
KLGM63437
Chicken src kinase, CSKM85039
EphB5U23783
EphA4*Z19059
Axl-related receptor tyrosine kinase, c-RekU70045
Chicken c-Eyk proto-oncogene (Axl/Tyro3/Mer family member)*L21719
Genes enriched in E4.5 DRG SSH-PCR subtraction library Fold Enrichment Compared to E8.5 DRG
GG ESTBU216199>3–5
Neural Epidermal Growth Factor-Like Like 2, NELL2D86747>3–5
Goliath 1-related zinc finger protein, G1rp (mouse)AF171875>1–3
GG ESTBU435181>1–3
StathminX65458>1–3
Lactate dehydrogenase H subunit, LDH-BAF069771>1–3
NovelIn progress>1–3
Y-box binding protein, YB-1L13032>1–3
GG EST, putative RING finger protein (human AJ009771)BU396906>1–3
GG ESTBU441482None
GG ESTBU448164None
HyperionAJ131892None
GG ESTBU262644None
GG ESTBU394286None
Per1 interacting protein of the suprachiasmatic nucleus, PIPS (rat)AB051807None
26S ATPase complex subunit 4U60187None
14-3-3zeta/tyrosine 3-monooxygenase tryptophan 5-monooxygenase activation protein zeta polypeptide (YWHAZ)AF465638None
GG ESTBX261356None
Homo sapiens KIAA0140 gene productNM_014661ND
GG ESTBU199356ND
GG ESTBY681404ND
Genes enriched in E8.5 DRG SSH-PCR subtraction library Fold Enrichment Compared to E4.5 DRG
  • *

    Identified as more robustly expressed in E4.5 DRG than in E8.5 DRG.

  • a

    DRG, dorsal root ganglia; E, embryonic day; PCR, polymerase chain reaction; RT, reverse transcriptase; SSH, suppressive subtractive hybridization; EST, expressed sequence tag.

NeuroserpinZ71930>1–3
Beta-III TubulinJ00913None
GG ESTBU477525None
GG ESTBU445425None
GG ESTBU273217None
cytochrome C oxidase subunit IIIX53434None
dynein light chain 2 (Dlc2, human)NM_080677None
GG ESTBU251703None
GG ESTBX269302None

Identification of Receptor Tyrosine Kinase Gene Family Members in the Immature DRG (E4.5)

Receptor Tyrosine Kinases (RTKs) regulate such fundamental cellular events as proliferation, differentiation, and survival, and have been shown to be necessary for normal development (Schlessinger, 2000). To identify additional RTKs expressed during DRG development, we used primers to conserved regions of RTKs and conducted an RT-PCR analysis with mRNA isolated from immature E4.5 DRG (Brandli and Kirschner, 1995). Our results are summarized in Table 1. Two differentially expressed RTKs cDNAs were identified by differentially screening E4.5 degenerate RTK cDNAs with E8.5 DRG degenerate RTK DIG-labeled probes (not shown). Sequencing identified these cDNAs as EphA4, a member of the Eph family of RTKs, and c-Eyk a member of the Axl/Tyro3 family of RTKs whose function is unknown. Although not differentially expressed, we also identified another member of the Axl/Tyro3 family of RTKs, c-Rek, which was expressed in both immature and mature DRG. In situ hybridization for c-Eyk, and immunocytochemistry for c-Rek confirmed their prominent expression in the immature E4.5 DRG (Fig. 1). Eyk is expressed in the inner core of the ganglion where postmitotic neurons are localized and hence is not expressed by the mitotically active progenitor cells that encompass the perimeter of the DRG. In contrast, Rek is expressed both in the neuronal and progenitor cell regions of the immature DRG.

thumbnail image

Figure 1. Two RTKs identified by RT-PCR are prominently expressed in the immature DRG. A: E4.5 sections from forelimb levels were hybridized with in situ probes for c-eyk. Note expression in the DRG (arrows), spinal cord (sc), sympathetic ganglia (sg), and floor plate. B: E4.5 sections were labeled with an antibody to Rek; note strong expression throughout the DRG.

Download figure to PowerPoint

Subtraction Analysis

To search for new genes preferentially expressed in the immature DRG, we directly compared the E4.5 library to cDNA generated from E8.5 DRG with a suppressive subtractive hybridization PCR (SSH-PCR) / differential screening methodology (Lukyanov et al., 1995; Munroe et al., 1995; Diatchenko et al., 1996; von Stein et al., 1997). We differentially screened approximately 600 clones from our E4.5 subtracted library with probes derived from total, unsubtracted cDNAs from both ages, and the subtracted cDNAs from both ages. Figure 2 shows the typical results of our differential screening. The top long arrow points to an amplicon from a clone (identified later as neural epidermal growth-factor like-like 2, NELL2) having different signal intensities with these probes, indicating that SSH-PCR has significantly enriched NELL2 at E4.5 and that its expression has decreased by E8.5. The bottom row is a series of control amplicons, and the lower right arrow points out the subtraction of the glyceraldehyde 3-phosphodehydrogenase (G3PDH) amplicon in both subtracted probe sets, while strong signals are seen in both unsubtracted probe sets, indicative of the success of SSH-PCR. The amplicon in the lower left corner (left arrow) is a random clone picked from the E8.5 subtracted library as a reverse control, showing that it was significantly enriched by the reverse experiment and, interestingly, was later identified as Neuroserpin, a gene known to be involved in later stages of neural differentiation (Osterwalder et al., 1996; Krueger et al., 1997). Amplicons with signal intensities similar to that of the clone identified later as NELL2 were selected for sequencing analysis. We also randomly picked additional clones from both the E4.5 and E8.5 subtraction libraries for further analysis. A list of the isolated cDNAs identified in the E4.5 subtracted and unsubtracted library is summarized in Table 1.

thumbnail image

Figure 2. Differential screening. Films from quadruplicate arrays of amplified cDNAs from the E4.5 DRG subtraction library hybridized with E4.5 subtracted and unsubtracted, and E8.5 subtracted and unsubtracted P32 labeled probes. Top right long arrow indicates a clone identified as NELL2. Bottom right short arrow is the subtraction control G3PDH. Bottom left short arrow indicates a clone identified as Neuroserpin.

Download figure to PowerPoint

Confirmation of Differential Gene Expression by Quantiative PCR

To verify that the cDNAs isolated in our screens were differentially expressed, we used Quantitative PCR (QPCR) to compare the relative expression levels of the identified cDNAs from the differential-screen. Of the 21 cDNAs identified in our screen as being more strongly expressed in the immature DRG than mature DRG, 9 differential screening patterns were validated by QPCR (Table 1). Of these cDNAs, one is a secreted glycoprotein: NELL2 (Matsuhashi et al., 1995; Nelson et al., 2002), one is a transcription factor: Y-box binding protein YB-1 (Kuwano et al., 2003), two are putative RING finger proteins: Goliath-1 and an EST corresponding to an unknown gene, one is a completely novel sequence, two correspond to ESTs that correspond to unknown genes, and one is Stathmin, a growth associated protein whose expression decreases in adult brain (Mori and Morii, 2002). As expected the majority of the randomly picked clones were not confirmed to be differentially expressed, while the clones selected on the basis of their differential signal intensities were validated. This is not surprising as the complexity of libraries generated by SSH-PCR are influenced by several factors, but primarily by the relative starting concentrations of cDNA fragments in the respective tester and driver tissues, and the degree of similarity of the two tissues being compared (Ji et al., 2002). We are now in the process of isolating full-length cDNAs from our full length E4.5 phage library.

Confirmation of Gene Expression in the DRG by In Situ Hybridization

We next conducted in situ hybridizations on sections of E4.5 and E8.5 embryos with probes generated from all of the isolated cDNAs (Figs. 3, 4). As the size of the cDNA fragments isolated from our screens varied from ∼100–800 bp, with most ∼200–300 bp, not all of the subtraction clones worked well for in situ hybridizations. To determine whether subtraction clones were expressed in progenitor cells or in postmitotic neurons, E4.5 embryos were injected with a 2-hr pulse of BrdU before fixation, and sections were triple labeled with antibodies to neurofilament, BrdU and counterstained with DAPI. At this stage, there is a spatial organization to the DRG such that postmitotic neurons reside in the interior region of the ganglion and are ringed by a perimeter of mitotically active progenitor cells (Nelson et al., 2002). The dorsal pole of the ganglion consists exclusively of mitotically active progenitor cells and the centrally projecting axons of sensory neurons. By E8.5 this spatial organization is no longer present and instead neurons are distributed throughout the ganglion (Oakley et al., 1997; Rifkin et al., 2000). Note that the neurofilament antibody used (Zymed RM0 270) labels all neurons at E4.5 but only the larger diameter neuronal cell bodies in the E8.5 DRG: it does not label the small diameter neuronal somata in the DRG nor sympathetic somata although it does label their axons. Because the small diameter neurons occupy the dorsal medial portion of the ganglion, the lack of NF labeling of that region should not be inferred to mean the absence of neurons—in fact that region is densely packed with small diameter neurons (Hamburger et al., 1981; Lefcort et al., 1996; Oakley et al., 1997).

thumbnail image

Figure 3. Both NELL2 and an EST BU396906, encoding a putative RING Finger protein, are more strongly expressed in the immature E4.5 DRG than in the mature E8.5 DRG. Forelimb level sections through both E4.5 (a–d,h–k) and E8.5 (e–g,l–n) embryos were hybridized with probes to either NELL2 (a–g) or BU396906 (h–n) and triple labeled with antibodies to BrdU (E4.5 embryos only, see Experimental Procedures, c,d,j,k) and neurofilament (b,d,f,g,I,k,m,n) and DAPI 9d,g,k,n).

Download figure to PowerPoint

thumbnail image

Figure 4. Additional cDNAs identified in the screen are expressed in the DRG. E4.5 and E8.5 sections were hybridized with probes for Cytochrome C oxidase III (a–g), 26S ATPase (S4; h–n), and Dynein light chain 2 (o–r), and labeled with antibodies as above: BrdU (c,d,j,k,q,r), neurofilament (b–d, f,g,i–k,m,n,p–r,t,u).

Download figure to PowerPoint

Figure 3 reconfirms the SSH-PCR and QPCR demonstrating the differential expression pattern of two of the cDNAs: NELL2 and an EST corresponding to an unknown gene. Both are more strongly expressed in the immature DRG than in the mature DRG. Additional combined in situ hybridization/immunocytochemistry for NELL2 revealed that NELL2 expression in a subset of mitotically active progenitor cells decreased over time (∼8% in rostral DRG and 25% in caudal DRG at E3.5; ∼4% in rostral DRG at E4.5; not shown), suggesting that NELL2 may be involved in neuronal differentiation. Our functional data confirms NELL2's role in neuronal development (Nelson et al., submitted). The uncharacterized EST BU396906, a putative Ring finger protein, also exhibits an interesting expression pattern whereby it is expressed both in nascent neurons in the core of the ganglia, and in mitotically active progenitors in the dorsal pole while it is excluded from the ventral pole and peripheral nerve.

Figure 4 demonstrates nervous system specific expression of three other cDNAs corresponding to identified genes, 26S ATPase cS4, dynein light chain 2, and Cytochrome C Oxidase III, that were expressed in the DRG at both ages. At E4.5, all three were expressed in both the progenitor and neuronal zones, although 26S ATPase cS4 and dynein light chain 2 were more strongly expressed in postmitotic neurons. In the spinal cord, while predominant expression of dynein light chain 2 is observed in the postmitotic differentiating mantle zone, its expression overlaps that of the most lateral region of the mitotically active ventricular zone. By E8.5, all three were expressed by the vast majority of neurons in the DRG. It will be of interest to determine the function of these genes during DRG development.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

A powerful strategy for understanding the molecular mechanisms governing the development of specialized cells is to identify the mRNA transcripts that are selectively expressed at discrete stages. Toward this end we have used two methods for identifying genes expressed during the peak period of sensory neurogenesis in the embryonic DRG. One of these approaches was designed to identify members of the receptor tyrosine kinase gene family known to exert potent activities in many developing systems, whereas the second approach, suppressive subtractive hybridization, was designed to identify a set of transcripts expressed specifically during the peak period of neurogenesis in the DRG.

Both approaches have identified transcripts for genes that are known to play significant roles in the developing nervous system. Using primers to conserved regions of RTKs, c-Rek, c-eyk, EphA4 and EphB5, CSK, PDGFa-1 receptor, and KLG were identified as being expressed in the immature E4.5 DRG (Table 1). Both c-Rek and c-eyk are members of the Axl/Tyro3 family of RTKs and their mammalian homologues have been shown to be prominently expressed in the developing nervous system (Lai et al., 1994). The expression of rek is developmentally regulated, being maximally expressed in the retina during the period of neuronal and glial cell differentiation (Biscardi et al., 1996; Fiordalisi and Maness, 1999). We show here that both c-Rek and c-Eyk were strongly expressed in the DRG during the peak period of neurogenesis and differentiation (Fig. 1). Activated c-Eyk constitutively stimulates the JAK-STAT pathway while only moderately activating the Ras-MAP kinase and JNK pathways (Zong et al., 1996; Besser et al., 1999). It has two C2-type immunoglobulin domains and two FN-III domains in its extracellular region, suggesting that it may play a role in cell–cell interactions (Jia and Hanafusa, 1994). Its normal cellular function is not known and is currently under study. EphA4 and EphB5, members of the Eph family, play significant roles during motor axon guidance (EphA4) in addition to influencing other developing systems (Sajjadi and Pasquale, 1993; Soans et al., 1996; Eberhart et al., 2000; Helmbacher et al., 2000). CSK (C-terminal Src kinase) down-regulates Src activity by phosphorylating a key inhibitory tyrosine, and is strongly expressed in developing brain followed by a dramatic down regulation in the adult brain (Kuo et al., 1997). A homologue, CHK (CSK homologous kinase) has been shown to directly associate with TrkA upon NGF stimulation and is involved in neurite outgrowth (Yamashita et al., 1999). PDGFα-1 receptor is involved in both sensory and motor neuron development (Li et al., 1996). KLG is a catalytically inactive member of the RTK superfamily with several immunoglobulin domains extracellularly, is homologous to CCK-4 (colon carcinoma kinase), and has been hypothesized to form a complex with a kinase-active receptor, reminiscent of the erbB3:erbB2/B4 interactions (Chou and Hayman, 1991; Mossie et al., 1995; Miller and Steele, 2000).

Evidence for the effectiveness of our suppressive subtractive hybridization (SSH-PCR) screen includes: (1) G3PDH, a highly abundant house-keeping gene present in both E4.5 and E8.5 transcriptomes was efficiently subtracted out and that (2) clones from the E4.5 subtraction library show differential signal intensities indicating that enrichment occurred for a subset of E4.5 specific genes in the forward experiment. This demonstrates that our SSH-PCR analysis effectively enriched differentially expressed transcripts and subtracted common, housekeeping transcripts. We also confirmed by QPCR that the cDNAs selected on the basis of their differential signal intensities from the forward subtraction were in fact differentially expressed in the immature E4.5 DRG compared with mature E8.5 DRG, in addition to confirmation by in situ hybridization that two of these cDNAs, NELL2 and an unknown EST are more robustly expressed in immature E4.5 DRG than in the mature E8.5 DRG (Fig. 3 and Nelson et al., 2002). The utility of this strategy is also supported by the fact that the functions of the identified gene products correlate with mediators of important cellular events during development. For example, in ongoing work we have evidence that NELL2 induces neuronal differentiation within the DRG (Nelson et al., submitted). NELL2 belongs to an intriguing class of proteins that contain EGF-like domains (Watanabe et al., 1996; Engel, 1989; Davis, 1990), including amongst others, Notch and its ligands Delta/Serrate/Lag (DSL), the Neuregulin (NRG) family of ligands, and the SLIT family of ligands. NELL2 is a secreted glycoprotein and belongs to the Laminin G/N-terminus Thrombospondin 1 (N-TSP1)/Pentraxin gene superfamily (Beckmann et al., 1998). Recently Noelin1, another secreted glycoprotein, has been shown to function during neural crest formation and migration, underscoring the importance of this molecular class in key developmental processes (Barembaum et al., 2000). Based on the subtractive screen, we conducted an in depth analysis of NELL2 expression in the developing nervous system and functional studies to determine its role during DRG development, thus demonstrating the feasibility of this approach and strategy for identifying the function of the remaining candidate genes from our analysis (Nelson et al., 2002; Nelson et al., submitted).

Based on their sequence, many of the cDNAs identified in the screen are potentially interesting candidate genes whose function within the DRG will be worth elucidating. These include genes that have been associated with regulation of the cell cycle and cell proliferation including YB-1 and stathmin (Mori and Morii, 2002; Kuwano et al., 2003; Bieche et al., 2003). Intriguingly, two of the differentially expressed cDNAs are RING finger proteins which in many systems have been shown to function as ubiquitin-protein ligases and to mediate major developmental events including Notch activation, by regulating cell surface levels of Delta expression and the generation of anterior ectoderm (Deblandre et al., 2001; Borchers et al., 2002; Hatakeyama and Nakayama, 2003). Several of the cDNAs identified in the screen but not confirmed by QPCR to be differentially expressed, also belong to intriguing gene families with proven functions in the developing nervous system. These include another protein involved in selective protein degradation, ATPase complex subunit 4 (S4) one of several different subunits of the multimeric regulatory complex that combines with the 20S proteolytic core to form the 26S proteasome (Dubiel et al., 1992; Singh et al., 1996, Hutson et al., 1997). Regulation of cellular proliferation through ubiquitin-mediated proteolytic degradation of cell cycle components is a crucial function of the proteasome, and its levels are particularly high in immature and mitotically active cells (Ichihara and Tanaka, 1995). Furthermore, selective degradation is as essential as the selective transcription/translation events normally associated with these cellular events. None of these three genes had been previously identified in the DRG and their expression during the peak of neurogenesis underlies the utility of this screen in identifying genes that potentially play key roles in the developing nervous system. Another molecule identified in our screen is the zeta isoform of the 14-3-3 family of proteins, which are key regulators and integrators of many different signal transduction pathways, and mediate such diverse cellular processes as proliferation, differentiation, survival, and apoptosis (Skoulakis and Davis, 1998; Fu et al., 2000). Of interest, they are abundantly expressed in the nervous system and disruptions in their function perturb neuronal differentiation (Chang and Rubin, 1997; Kockel et al., 1997; Li et al., 1996; Skoulakis and Davis, 1998). For example, disruption of 14-3-3 binding to the B-Raf kinase domain uncouples NGF induced cellular differentiation of PC12 cells (MacNicol et al., 2000). Further experimentation will be necessary to elucidate the function of these genes during DRG development.

In summary, our analysis has successfully identified genes likely to be involved in mediating neurogenesis and differentiation. The known and proposed functions of the identified proteins correlate well with the developmental stages of the DRG in which they are expressed. Investigations into their function should provide considerable insight into the cellular and molecular mechanisms regulating sensory neurogenesis and differentiation in the DRG. With the completion of a draft sequence of the chicken genome in 2004, the compilation of several comprehensive EST resources, and the eventual generation of chick microarrays, combined with the ease of gene manipulation in the chick, the classic embryologic chick model system returns to the forefront of developmental biology (Swartz et al., 2001; Boardman et al., 2002; Burt and Pourquie, 2003).

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Construction of Full-Length Phage E4.5 Chick DRG cDNA Library

Fertilized white leghorn eggs were obtained (Spafas), and grown to appropriate stages (Hamburger and Hamilton, 1951). All embryos were treated in accordance with IUCAC stipulations. mRNA from E4.5 DRG dissected from all axial levels was isolated by using MicroPoly(A)Pure (Ambion, Inc.) as directed. This was converted to full-length double-stranded cDNA and packaged into Lambda phage as directed resulting in a phage titer of ∼1.0 × 108 plaque forming units (pfu)/ml (Stratagene). PCR-DIG labeled probes generated from a fragment of alpha-tubulin and the extra-cellular domain of TrkC were used to screen the library (see below), candidates were selected, inserts checked by means of restriction analysis, sequenced, and identified by blasting GenBank (Altschul et al., 1990).

Receptor Tyrosine Kinase Gene Family RT-PCR

Fertilized white leghorns were obtained (Spafas), and grown to appropriate stages (Hamburger and Hamilton, 1951). All embryos were treated in accordance with IACUC stipulations. RT-PCR was used to analyze the E4.5 DRG transcriptome for the presence of receptor tyrosine kinase (RTK) gene family members with a degenerate set of primers (Brandli and Kirschner, 1995; see below for primer sequences). mRNA from E4.5 DRG from all axial levels was isolated by using MicroPoly(A)Pure (Ambion, Inc.) as directed, and single-stranded cDNA synthesized with an oligo-dT primer and AMV-RT (Invitrogen), 42°C for 1 hr. All PCR experiments were conducted with GeneAmp 2400/9600 machines (Perkin Elmer). Touchdown PCR (anneal 60°C for cycles 1–5X, 55°C for 6–10X, 50°C for 11–40X) amplifications with this primer set were analyzed on 2% Ag/Br gels, products gel extracted (QIAquick Gel Extraction, Qiagen), and TA-cloned into PCRII (Invitrogen). Plasmids were isolated and sequenced (Iowa State DNA Sequencing Facility). For identification, sequences were aligned to Genbank by using the BLAST program.

Suppressive Subtractive Hybridization PCR (SSH-PCR)

SSH-PCR was used to analyze transcriptome differences between immature, embryonic day 4.5 dorsal root ganglia (E4.5 DRG) and mature, fully differentiated E8.5 DRG (Lukyanov et al., 1995; Diatchenko et al., 1996). mRNA was isolated from E4.5 and E8.5 DRG from all axial levels using MicroPoly(A)Pure (Ambion, Inc.), as directed. These mRNAs (2 μg each) were then prepared for SSH-PCR analysis by using PCR-Select (Clontech), as directed. The forward subtraction experiment consisted of subtracting E8.5 DRG cDNA (driver) from E4.5 DRG cDNA (tester), enriching for E4.5 specific cDNAs. The reverse subtraction experiment consisted of subtracting E4.5 DRG cDNA (driver) from E8.5 DRG cDNA (tester), enriching for E8.5 specific cDNAs. E4.5 and E8.5 DRG cDNA subtraction libraries were created from both the forward and reverse subtraction experiments by TA-cloning an aliquot of each secondary, nested PCR reaction into the PCR II vector (Invitrogen). A modified secondary, nested PCR reaction was used to radiolabel (dCTP32 800 Ci/mMol, New England Nuclear) or chemically label (PCR DIG dNTP labeling mix, Roche Molecular Biochemicals) E4.5 and E8.5 DRG subtracted and unsubtracted control cDNAs for differential screening. Unincorporated label was removed from all probes with QIAquick PCR Purification spin columns (Qiagen).

DRG cDNA Subtraction Library Differential Screening

The E4.5 DRG cDNA subtraction library was differentially screened for candidate differentially expressed clones (Munroe et al., 1995; von Stein et al., 1997). Initially 300 random, recombinant clones were selected for differential screening and used to inoculate 100 μl LB-kan/amp cultures grown overnight at 37°C. Inserts were amplified by colony PCR with the nested primer set, and analyzed on large format, 2% Ag/EtBr gels. The remaining cultures were then mixed with glycerol and frozen at −80°C. A total of 0.5 μl of each diluted amplicon (1:10) was arrayed in duplicate, onto quadruplicate, positively charged nylon membranes (Osmonics, Inc.). Because the nested primer set was used to create both the amplicons and the probes, oligos corresponding to the sequences of both adaptors were synthesized (MWG Biotech, Inc.) and used to block hybridization signals due to adaptor background. As a control for adaptor background, the 603-bp band from the HaeIII digest of the Phi-X174 phage DNA/(Promega) control subtraction experiment (PCR-Select, Clontech), which was flanked by both adaptors, was gel extracted (QIAquick Gel Extraction, Qiagen), re-amplified with the nested primer set, and prepared for arraying. As a control for subtraction efficiency, specific primers were used to amplify a segment of G3PDH, which was then prepared for arraying (see RT-PCR methods for sequences). As further controls for subtraction efficiency, five random, recombinant clones from the E8.5 DRG cDNA subtraction library were also prepared for arraying. Membranes were denatured in 0.5 M NaOH, 1.5 M NaCl for 2 min, neutralized in 0.5 M Tris-Cl, pH 7.5, 1.5 M NaCl for 5 min, rinsed in 0.2 M Tris-Cl/2× SSC for 30 sec, and UV cross-linked (Stratalinker, Stratagene). Membranes were prehybridized with the blocking oligos for 1 hr in standard hybridization buffer at 72°C in a Mini Oven MKII (Hybaid). Probes were mixed with another equivalent of blocking oligos and denatured by boiling for 5 min, cooled on ice for 2 min, added to the membranes and hybridized overnight at 72°C. Membranes were washed twice in 2× SSC/0.1%(w/v) SDS for 15 min at 72°C, and then washed twice again in 0.5× SSC/0.1% (w/v) SDS for 15 min at 72°C. Radiolabeled membranes were then exposed to film (BioMax, Kodak) for 2–7 days at −80°C. DIG labeled membranes were prepared for chemiluminescent detection as directed (Roche Molecular Biochemicals) and exposed to film (BioMax, Kodak) for 5–15 min at room temp. Candidate differentially expressed cDNA clones were grown, plasmids were isolated with the QIAprep mini-prep kit (Qiagen), sequenced on an ABI PRISM Genetic Analyzer with the BIG DYE Terminator kit, and analyzed with Sequencher 3.0 (Gene Codes). Candidates with signals specifically enriched by the subtraction were picked for identification; sequences were aligned to GenBank by using the BLAST program. From the first 300 clones arrayed, ∼40 were chosen for sequencing. This yielded a large number of cDNA inserts corresponding to NELL2. This NELL2 cDNA fragment was used as a probe to re-screen the E4.5 subtraction library. Three hundred more non-NELL2 clones were arrayed and re-screened as above, which resulted in the identification of additional cDNAs exhibiting differential signal intensities. Furthermore, additional candidate cDNAs were randomly picked and sequenced from both the E4.5 DRG and E8.5 DRG SSH-PCR subtraction libraries.

Confirmation of Differential Gene Expression

Quantitative RT-PCR (QPCR) was used to confirm differential gene expression. E4.5 and E8.5 DRG were collected as described above, and total RNA isolated with standard techniques. Genomic DNA contamination was removed by digesting with 2U RQ1 RNase-free DNase (Promega) for 30 min at 37°C, phenol/chloroform extracting, and isoproponol precipitation. Total RNA was resuspended in 20 μl nuclease-free H2O, and integrity and concentration analyzed by capillary gel electrophoresis. Primer 3.0 was used to design specific primer sets from available sequence data for candidate chicken clones. Equal concentrations of E4.5 and E8.5 DRG total RNA were used to generate oligo-dT primed cDNAs with SuperScriptII (Invitrogen), and RT(-) controls. cDNAs were diluted in a series and used as template for QPCR with GAPDH primer set to establish the correct dilution for real-time analysis by using the SYBR Green QPCR Master Mix reagent (Applied Biosystems) and Opticon DNA Engine real-time QPCR machine (MJ Research). It was determined that a cDNA dilution of 1:12 gave optimal, reproducible results for GAPDH ∼16 cycles, E4.5 and E8.5 DRG cDNA sample concentrations were equal, and the RT(-) controls did not contain genomic DNA. Master mixes were assembled for each respective cDNA template (including RT(-) controls) and SYBR Green PCR mix. Each candidate primer set was run in triplicate with one additional sample for RT(-) control for both E4.5 and E8.5 cDNA templates. Master mixes of each primer set were assembled and mixed with the cDNA-SYBR reagent master mix, and cycled through a standard Cycle program (94°C 10 min, and 42 cycles of 94°C 30 sec, 60°C 30 sec, 72°C 1 min, and melting curve analysis. Samples were normalized relative to the average of three C(T) (cycle at which amplicon reaches log phase of amplification). Each cycle number differences between respective samples compared with GAPDH levels represent a twofold change, and are reported as the fold difference between respective samples at E4.5 and E8.5 DRG cDNA.

In Situ Hybridizations and Immunocytochemistry

Triple labeling of paraffin sections was performed as described (Nelson et al., 2002). Briefly embryos were incubated to E4.5 and E8.5. E4.5 embryos were windowed, pulsed with BrdU, resealed, and incubated for 3–4 hr to label mitotically active progenitor cells. Embryos were collected, fixed in modified Carnoy's solution (60% EtOH/30% formaldehyde/10% acetic acid) overnight at 4°C, dehydrated though an EtOH series, and prepared for paraffin embedding and sectioning; sections were collected from the level of the forelimb. In situ hybridizations were performed as described (Etchevers et al., 2001; Nelson et al., 2002). DIG-labeled anti-sense RNA probes were in vitro transcribed from linearized E4.5 SSH-PCR clone corresponding to EST BU396906, full-length NELL2 plasmid (Nelson et al., 2002), E8.5 SSH-PCR clones corresponding to cytochrome C oxidase subunit III (X53434), and ESTs BU477525 and BU445425, and GEISHA (Gallus gallus [chicken] EST and In Situ Hybridization Analysis database) chicken cDNA clones corresponding to dynein light chain 2 (29M, X79088) and YB-1 (08E, L13032): GEISHA database at www.geisha.biosci.arizona.edu, (gift of Parker Antin). All other clones were also used for in situ hybridizations, but due to the small size of these clones (∼100–300) not all of them worked well for in situ hbridizations. After successful in situ hybridizations, sections were fixed in 4% paraformaldehyde, rinsed in PBS, acid treated for 8 min in 4 N HCl, washed 4× in PBT (PBS-0.1% Triton X-100), blocked for 1 hr in 10% goat serum PBT at room temp. Primary antibodies were diluted in block and incubated overnight at room temp: rat anti-BrdU 1:200 (Accurate Biochemicals); mouse anti-neurofilament 1:3,000 (Zymed RMO 270). Slides were washed and incubated with corresponding goat secondary antibodies at 1:500 for 1 hr at room temp; goat anti-rat:Alexa 566 and goat anti-mouse:Alexa 488 (Molecular Probes). Slides were washed and counterstained with Dapi and mounted. Images were acquired on an Axiovert Flourescent microscope (Zeiss) equipped with Normarski optics and a Spot camera. Images were compiled with Adobe Photoshop 6.0.

GenBank Submission

Chicken 14-3-3z partial cDNA sequence has been submitted to GenBank, accession number AF465638.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Dr. Patricia Maness for the c-Rek antibody and Dr. Daniel Besser for the c-Eyk cDNA plasmid, Valerie Todd for excellent technical assistance, Dr. Valerie Copie for sequencing support, Dr. Roger Bradley for helpful advice on in situ hybridizations and comments on the manuscript, Dr. Parker Antin and Geisha for cDNAs, and Dr. Thomas Reh for support of B.R.N. F.L. and B.R.N. were funded by the NIH and B.R.N. was funded by the Kopriva Foundation.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES