SEARCH

SEARCH BY CITATION

Keywords:

  • Xenopus laevis;
  • lens;
  • retina;
  • cornea;
  • neuronal leucine rich repeat protein;
  • morpholino

Abstract

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

Leucine-rich repeat proteins expressed in the developing vertebrate nervous system comprise a complex, multifamily group, and little is known of their developmental function in vivo. We have identified a novel member of this group in Xenopus laevis, XlNLRR-6, and through sequence and phylogenetic analysis, have placed it within a defined family of vertebrate neuronal leucine-rich repeat proteins (NLRR). XlNLRR-6 is expressed in the developing nervous system and tissues of the eye beginning at the neural plate stage, and expression continues throughout embryonic and larval development. Using antisense morpholino oligonucleotide (MO) -mediated knockdown of XlNLRR-6, we demonstrate that this protein is critical for development of the lens, retina, and cornea. Reciprocal transplantation of presumptive lens ectoderm between MO-treated and untreated embryos demonstrate that XlNLRR-6 plays autonomous roles in the development of both the lens and retina. These findings represent the first in vivo functional analysis of an NLRR family protein and establish a role for this protein during late differentiation of tissues in the developing eye. Developmental Dynamics 235:1027–1041, 2006. © 2006 Wiley-Liss, Inc.


INTRODUCTION

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

Leucine-rich repeats (LRR) are found in a complex and diverse superfamily of proteins, including those involved in early embryonic patterning (Hashimoto et al.,1988), brain development, and axon pathfinding (Rothberg et al.,1988; Nose et al.,1992; Taniguchi et al.,2000; Fukamachi et al.,2001), nervous system regeneration (Ishii et al.,1996; Bormann et al.,1999), learning and memory (Bando et al.,2005), and various mammalian cancers (Almeida et al.,1998; Hamano et al.,2004). Whereas LRR-containing proteins vary greatly in their number of LRR, tertiary structure, subcellular localization, enzymatic and signaling activities, they share a repeated sequence of 20–29 amino acids rich in stereotypically placed leucines and other aliphatic residues (Kobe and Deisenhofer,1994, 1995; Hayata et al.,1998).

With the resolution of the crystal structure of the LRR-containing porcine ribonuclease inhibitor, Kobe and Deisenhofer (1994, 1995) established that each LRR forms a hydrophobic β-sheet with conserved aliphatic residues followed by a more variable, parallel α-helix. Serial placement of these β- units forms a concave, horseshoe-shaped structure with a large surface area optimal for protein–protein interactions at the adjacent β-sheets. Subsequently to this characterization, LRR-containing proteins have been implicated in protein–protein interactions such as adhesion and intercellular signaling (Taniguchi et al.,1996; Fukamachi et al.,2001, 2002).

A family of LRR-containing proteins was first identified in the mouse as having a characteristic 11–12 LRR, flanked at the N-terminal and C-terminal sides with stereotypically placed cysteine residues (amino flanking region, NFR; carboxy flanking region, CFR), and also containing an immunoglobulin type C2 (IgC2) loop and a single transmembrane domain with a very short intracellular carboxy-terminal region. Some of these proteins also contain a fibronectin type 3 (FnIII) domain, an RGD integrin binding motif, and a clathrin-mediated endocytosis motif (see Fig. 1A). These proteins were named neuronal leucine-rich repeat (NLRR) proteins, based on their recovery from a neonatal mouse brain cDNA library (Taguchi et al.,1996). Members of the NLRR protein family have since been identified in the frog Xenopus, zebrafish, rat, and human, with predicted protein homologs in many other vertebrates. See Table 1 for selected members of this family.

thumbnail image

Figure 1. Characteristic neuronal leucine-rich repeat proteins (NLRR) protein features and inferred XlNLRR-6 protein sequence. A: The schematic diagram illustrates the conserved features among the known NLRR proteins. Not all members of the NLRR family have every illustrated domain; see Table 1. N-terminal (extracellular) domains are toward the left; C-terminal domains are toward the right. S, signal peptide; NFR, amino-flanking region with stereotypically placed cysteine residues; LRR, leucine-rich repeats; CFR, carboxy-flanking region with stereotypically placed cysteine residues; IgC2, immunoglobulin type C2 domain; FnIII, fibronectin type III domain; T, transmembrane helix. Also indicated are the positions of the RGD sequence in the eighth LRR of NLRR-3 proteins (black arrowhead) and the clathrin-mediated endocytosis motif of NLRR-1, -3, and -6 proteins (white arrowhead). B: The predicted protein sequence of XlNLRR-6 is shown. The LRR regions are aligned, illustrating the consensus sequence of xLxxLxxLxLxxNxL. In some cases, other aliphatic residues substitute for leucines (e.g., valine or isoleucine). Residues that differ from the consensus residues are indicated in lighter shading. Also indicated are the IgC2 domain (gray shading), transmembrane helix (double underline), and clathrin-mediated endocytosis motif (single underline).

Download figure to PowerPoint

Table 1. Features of NLRR Protein Family Members
GroupSpeciesLengthLRRIgC2FnIIIRGDEndoaAccessionReference
  1. See text and Figure 1A for further details. a, clathrin-mediated endocytosis motif; b, “hNLRR-2” was proposed as new nomenclature for the human glioma amplified on chromosome 1 protein (GAC1) by Hamano et al. (2004).

NLRR-1M. musculus716 aa12++D45913Taguchi et al.,1996
 X. laevis718 aa12+++AB014462Hayata et al.,1998
 H. sapiens716 aa11++BC034947Hamano et al.,2004
NLRR-2 ([TRIPLE BOND]GAC1)bM. musculus730 aa11+D49375Taguchi et al.,1996
 H. sapiens713 aa11+AY358290Hamano et al.,2004
NLRR-3M. musculus707 aa11++++D49802Taniguchi et al.,1996
 R. norvegicus707 aa11++++AF291437Fukamachi et al.,2001
 H. sapiens708 aa11++++AB060967Hamano et al.,2004
NLRR-4M. musculus735 aa8+BC056458Bando et al.,2005
NLRR-5H. sapiens713 aa11+CAI14096Hamano et al.,2004
proposed: NLRR-6D. rerio744 aa12++n/aBormann et al.,1999
 X. laevis739 aa12++DQ315790Wolfe et al.,2004; this study

Additional vertebrate proteins have been described, which share some of the NLRR protein features shown in Figure 1A but appear to be less closely related. These proteins have been variously named leucine-rich repeat neuronal (LRRN or LERN) and leucine-rich repeat transmembrane neuronal (LRRTN or LRRTM) proteins; members of these groups are also expressed in the nervous system (Carim-Todd et al.,2003; Lauren et al.,2003).

Whereas little in vivo functional examination has been conducted on NLRR proteins, their roles in nervous system development have been predicted largely based on expression data and in vitro studies. Mouse NLRR-1, -2, and -3 (see Table 1) are all expressed during nervous system development, and expression continues in discrete regions of the adult brain (Taguchi et al.,1996; Taniguchi et al.,1996). In Xenopus, NLRR-1 is expressed in the central nervous system (CNS) ventricular zone of the early tadpole, suggesting a role in neural layer formation within the brain and spinal cord (Hayata et al.,1998).

Rat NLRR-3 has been examined in vitro and is implicated in long-term potentiation of epidermal growth factor (EGF) signaling. NLRR-3 transcription in fibrosarcoma cells was up-regulated after EGF stimulation and also in the presence of the consitutively active small G-protein ras, while transcription decreased in the presence of mitogen-activated protein kinase (MAPK) inhibitors, suggesting that NLRR-3 transcription is downstream of EGF by means of a MAPK-dependent pathway (Fukamachi et al.,2001). Further examination revealed that NLRR-3 facilitates internalization of the EGF ligand/receptor complex, which continues to transduce intracellular signal after internalization (Fukamachi et al.,2002). The authors concluded that EGF ligand/receptor complexes cluster with NLRR-3 proteins, and the NLRR-3 cluster encourages the formation of a clathrin-coated vesicle by means of a nine-residue motif in its intracellular domain (YPPLI[N or S]LWE). Because three groups of NLRR proteins contain this same motif (see Table 1), the roles of these proteins, in part, also may involve internalization of signaling molecules.

In addition to roles in cell signaling and neural development, three NLRR proteins have been identified as playing a role in vertebrate neural regeneration. Mouse NLRR-3 transcription is up-regulated after cortical injury, specifically in layer III of the cortex, whereas NLRR-1 and -2 do not demonstrate this up-regulation. Thus, NLRR-3 may be associated with neuronal regeneration and reconstruction in the cerebral cortex (Ishii et al.,1996). In zebrafish, zfNLRR demonstrates prominent expression in the developing eye and CNS and expression is dramatically up-regulated in neurons whose axons regenerate after damage to the optic nerve and spinal cord (Bormann et al.,1999). A newly identified Xenopus expressed sequence tag (EST) tentatively named W006, with significant sequence similarity to other NLRR proteins, shows similar expression in the developing eye and brain, with up-regulation in the regenerating lens and hind limb (Wolfe et al.,2004). zfNLRR and W006 have a high sequence similarity to each other, indicating a possible shared role in development and regeneration (see below).

We have cloned the full length of the W006 gene and examined its relationship to other NLRR family members, as well as to other neuronally expressed LRR-containing proteins. Based on these analyses, W006 belongs in a new NLRR group with zfNLRR, which we have named NLRR-6. The timing and organ specificity of W006 expression has been characterized as beginning at neurulation, and larval expression is restricted to the developing lens, retina, and brain. As there is no previous in vivo work examining the role of NLRR proteins during development, we set out to establish the role of W006 in eye development, where the transcript is readily detectable in the lens and retina. Using antisense morpholino oligonucleotide-mediated knockdown of W006, we have established that W006 is necessary for proper differentiation of the lens and retina and that these roles are autonomous to each tissue.

RESULTS

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

Isolation and Identification of a Novel Xenopus NLRR and Its Relationship to the Vertebrate NLRR Protein Family

A subtracted cDNA library enriched for genes up-regulated during initial cornea–lens transdifferentiation events (Henry et al.,2002; Walter et al.,2004) was screened as previously described (Wolfe et al.,2004). Among the recovered clones was a putative member of the NLRR protein family (EST W006, NCBI accession no. CK815995; see Wolfe et al.,2004).

Clone W006 as recovered showed a partial overlap at the inferred amino acid level with the highly conserved LRR domains of several members of the NLRR protein family, but appeared to be incomplete at the 5′-end of the transcript. Successive rounds of 5′-rapid amplification of cDNA ends (RACE) revealed a cDNA 3,021 bp in length, with a short 5′-untranslated region (UTR), 2,217-bp open reading frame, and 655-bp 3′-UTR followed by a polyadenylation signal (NCBI accession no. DQ315790). Northern hybridization of stage 32–36 poly-A RNA revealed a single band at approximately 3.1 kb, corresponding to the length of the W006 clone assembled with its RACE fragments (data not shown).

The complete inferred amino acid sequence of 739 residues is shown in Figure 1B. Twelve LRR of 24–26 amino acids each are identified, showing the conserved residues of the repeat regions. Other features characteristic of NLRR proteins, including cysteine-rich amino- and carboxy-flanking regions of the LRR and an immunoglobulin C2 type domain, are also identified. The short carboxy-terminal sequence contains the nine-residue YPPLI(N or S)LWE clathrin-mediated endocytosis motif present in NLRR-1, NLRR-3, and zfNLRR (Fukamachi et al.,2002). W006 demonstrates no significant match to a fibronectin type III domain, which has been reported in xNLRR-1 and other NLRR family members (see Fig. 1A; Table 1).

To establish whether W006 represents a Xenopus member of a previously characterized NLRR protein group or a novel group, NCBI blastp searches were conducted. Table 2 shows amino acid similarities among all of the previously described NLRR proteins. Whereas all of the members examined demonstrated some homology to one another, there was much more significant alignment among members of the same NLRR group, as is shown by darker shading in the table (see Discussion section below).

Table 2. Amino Acid Percent Similarities Among Selected NLRR Proteins
inline image

Because the zebrafish protein zfNLRR does not convincingly match the other family members, it deserves categorization as a new class, here called NLRR-6. Additionally, W006 bears the strongest resemblance to zfNLRR-6 and may belong in this group as well. However, because the protein similarity between zfNLRR and W006 is 67%, additional support was sought for placing them together in a new group.

A search of GenBank revealed a variety of known and predicted vertebrate proteins, which all shared several LRR regions, a single transmembrane domain, and those with characterized expression patterns were expressed in the nervous system. Many of these proteins had been named by means of conventions that were not distinctive (e.g., neuronal leucine-rich repeat, NLRR; vs. leucine-rich repeat neuronal, LRRN; and leucine-rich repeat transmembrane neuronal, LRRTN), and in some cases there seemed to be significant similarities between differently named proteins across groups. Thirty-six vertebrate proteins from these various groups were selected, along with likely outgroups consisting of two putative NLRR proteins from Caenorhabditis elegans, slit protein from Xenopus and Drosophila, and capricious and tartan from Drosophila. Together, all 41 proteins were aligned using ClustalX software and the readily alignable, more highly conserved NFR and CFR from each was selected for analysis of the proteins' phylogenetic relationships.

Bayesian analysis was conducted on the aligned regions to separate the proteins into related orthologous groups (Huelsenbeck et al.,2001). The unrooted consensus tree generated from this analysis is shown in Figure 2A. Whereas some branches were not well resolved, three clear orthologous groups were identifiable from the tree: the extensive LRRN-1, -2, and -3 and LRRTN-3 and -4 proteins (Fig. 2A); the vertebrate NLRR and LRRN-5 proteins (see also Fig. 2B); and the LRRN-6A and -6D proteins (see also Fig. 2C). The mouse and rat proteins previously named NLRR-4 did not appear to join any of the ortholog groups, supporting the weaker sequence similarity data shown in Table 2 and suggesting that these proteins should be further examined and possibly renamed.

thumbnail image

Figure 2. Phylogenetic analysis of relatedness of vertebrate neuronally expressed leucine-rich repeats (LRR) -containing proteins. A: Radial tree demonstrating relatedness of 41 LRR-containing proteins based on the well-aligned regions up- and downstream of the LRR regions (NFR and CFR, respectively). Drosophila slit, tartan, and capricious; Xenopus slit; and two Caenorhabditis elegans putative NLRR proteins were included in the analysis as outgroups to the vertebrate neuronally expressed LRR-containing proteins. The asterisk in A indicates the position of XlNLRR-6 in the tree. B: Magnified area of the shaded inset neuronal LRR (NLRR) Family in A. C: Magnified area of the shaded inset LRRN-6 Group in A. Each protein is named first with a species abbreviation followed by a protein name or abbreviation. Species abbreviations: Bt, Bos taurus; Ce, Caenorabditis elegans; Cf, Canis familiaris; Dm, Drosophila melanogaster; Gg, Gallus gallus; Hs, Homo sapiens; Mm, Mus musculus; Pt, Pan troglodytes; Rn, Rattus norvegicus; Xl, Xenopus laevis; zf, zebrafish (Danio rerio). Protein abbreviations: NLRR, neuronal leucine-rich repeat; LRRN, leucine-rich repeat neuronal; LRRTN, leucine-rich repeat transmembrane neuronal; GAC-1, glioma amplified on chromosome 1; (p), predicted identity of protein based on database comparison by original depositor. The scale bars indicate number of residue substitutions per site per unit branch length. Underlined numbers at nodes indicate fraction of bootstrap support for that node.

Download figure to PowerPoint

The 15 proteins in the ortholog group NLRR Family shown in Figure 2B were selected for a more thorough phylogenetic examination and were much more universally alignable in ClustalX. With the exception of a short, highly variable amino acid sequence at the N- and C-terminals, these proteins were aligned with three outgroup proteins from the LRRN-6 Group and mouse NLRR-4. The consensus rooted phylogram generated by Bayesian analysis is shown in Figure 3. The NLRR-1, -3, and -6 proteins form a sister group to the NLRR-2 and -5 proteins. Furthermore, the grouping of the zebrafish NLRR-6 protein with the W006 protein suggests a close homology between these proteins. The sequence and phylogenetic analyses of the protein previously cloned as W006 support the adoption of a shared naming convention with zfNLRR and, henceforth, will be referred to as zfNLRR-6 and XlNLRR-6.

thumbnail image

Figure 3. The neuronal leucine-rich repeat proteins (NLRR) Family phylogram based on full-length protein alignment. Rooted phylogram of orthologs within the NLRR Family from Figure 2B, based on their full-length protein sequences excluding short stretches of unalignable N-terminal and C-terminal residues. Two members of the LRRN-6 Group and the mouse NLRR-4 sequence are included as outgroups to root the tree. The scale bar indicates number of residue substitutions per site per unit branch length. Underlined numbers at nodes indicate fraction of bootstrap support for that node.

Download figure to PowerPoint

XlNLRR-6 Is Expressed in the Developing Eye and Brain Beginning at Neurulation

RT-PCR was conducted to establish the timing of XlNLRR-6 expression during embryonic and early larval development. Time points examined included cleavage stages 1–3 and 4–6, gastrulation stages 10 and 11.5, neurulation stages 14–15, neural tube stages 19–22, tail bud stages 25–27, and larval stages 36–37 and 43–45 (stages of Nieuwkoop and Faber,1956). Expression was first detected at neurulation, and was maintained throughout the larval periods examined (Fig. 4A); this temporal expression pattern is similar to that reported for xNLRR-1 (Hayata et al.,1998). The spatial pattern of XlNLRR-6 transcript was then examined by in situ hybridization. Expression before stage 20 was generally below the level of detection (data not shown); however, significant expression in the developing eye, otic vesicle, and central nervous system was evident as early as stage 26 (Fig. 4B). This pattern later resolved into lens, retina, and brain expression (Fig. 4C,D). By larval stages 37–39, prominent expression continued throughout each vesicle of the brain, while expression in the eye was predominantly restricted to the retina (Fig. 4E,F).

thumbnail image

Figure 4. Developmental expression of XlNLRR-6. A: Reverse transcriptase-polymerase chain reaction (RT-PCR) of XlNLRR-6 reveals expression beginning at neurula stages 14–15, and continuing through the early larval period. Expected product size for the PCR was approximately 2.1 kb, as indicated by the bands shown. The portion of the gel below 1.5 kb has been cropped (and contained no extraneous bands). B–F: In situ hybridization of XlNLRR-6 transcript. B, lateral view of tail bud stage 26. C: Stage 29 lateral view. D: Stage 35 lateral view. E: Stage 37 dorsal view of head. F: Stage 39 lateral view of head. ey, eye; fb, forebrain; hb, hindbrain; mb, midbrain; oc, optic cup; ov, otic vesicle; re, retina; sc, spinal cord. Scale bars = 500 μm in B–G.

Download figure to PowerPoint

XlNLRR-6 Knockdown Disrupts Differentiation of the Lens and Retina in a Dose-Dependent Manner

To examine the role of XlNLRR-6 in brain and eye development, an antisense morpholino oligonucleotide (MO) was generated against the translation initiation site of the XlNLRR-6 transcript. The MO is a chemosynthetic oligonucleotide designed to hybridize to this site and sterically block binding of ribosomes to the transcript (Heasman,2002). Due to its synthetic qualities, the MO is not degraded by cellular machinery and perdures for at least several days (Heasman,2002). The XlNLRR-6 MO was tagged with a 3′-lissamine tracer, allowing visualization of MO-targeted tissues.

XlNLRR-6 MO was injected into one cell of Xenopus embryos at the two-cell stage. Other two-cell embryos were treated with a control MO, which does not target any known transcripts, to look for general toxicity that might be caused by the introduction of MO during development. Gross morphological defects on the injected side of XlNLRR-6 MO-treated specimens were not visible throughout early development, through early neurulation stages (data not shown). Through the early tail bud stages, no morphological defects in neural/retinal development were visible in sectioned embryos. Figure 5 shows a representative XlNLRR-6 MO-treated specimen at stage 25, demonstrating morphologically and histologically normal optic vesicle and neural tube development. Only after formation of the lens and elaboration of the neural retina and pigment epithelium were defects noted. Stages 35–38 were selected for scoring these samples, due to the prominent pigmentation that develops in the retinal pigment epithelium by this stage. When compared with their uninjected sides, XlNLRR-6 MO-treated larvae demonstrated gross defects in eye size, pigmentation, and lens formation in these MO-targeted tissues (Fig. 6). Defects varied in severity, from slight diminishment in size and pigmentation to disorganization or even loss of structures in the eye, and these defects were not evident in control MO-treated specimens at comparable doses (compare Fig. 6A,B with C–F).

thumbnail image

Figure 5. Examination of eye tissues at stage 25 after Xenopus laevis neuronal leucine-rich repeat proteins (XlNLRR) -6 knockdown. Example embryo injected with 9 ng of XlNLRR-6 morpholino oligonucleotide (MO) in a single cell at the two-cell stage. A: The left side of the embryo did not receive MO treatment. B: The right side of the embryo (asterisk in E) incorporated the MO. C,D: Red fluorescence images showing absence of the MO lissamine marker on the left (C) and presence on the right (D). E: A 10-μm section through the developing forebrain of the embryo shown in A–D demonstrates that both sides of the neural tube and both optic vesicles appear normal. This timing is approximately one stage before the morphological appearance of the lens vesicle (Schaefer et al.,1999). cg, cement gland; nt, neural tube lumen; ov, optic vesicles. Scale bar = 100 μm in E.

Download figure to PowerPoint

thumbnail image

Figure 6. Effects of Xenopus laevis neuronal leucine-rich repeat proteins (XlNLRR) -6 knockdown on lens and retina development. All specimens are shown at stages 35–38, after retinal pigment epithelium has formed in control embryos. Fluorescent images in the second row are the identical specimens to those in the first row; these show the distribution of the lissamine-tagged morpholino oligonucleotide (MO). White arrowheads show position of the eye, when present. A,G: Uninjected control specimen. B,H: A total of 9 ng of control MO were injected at the two-cell stage, leading to normal eye morphology. C–F,I–L: A total of 9 ng of XlNLRR-6 MO were injected at the two-cell stage. Typical eye defects from this treatment are shown in order of increasing severity. C,D,I,J: Small eye with progressive ventral developmental defects. E,K: Eye dysgenesis. F,L: Missing eye. Note that head size decreases as eye phenotype severity increases in C–F. M–R: Ten-micrometer sections at the level of the dashed line in A (mid-eye), stained with hematoxylin. The dorsal–ventral axis is from top-to-bottom, and the proximal–distal axis is from right-to-left in each section. M: Normal eye with visible lens differentiation and formation of distinct retinal layers. N–R: Eye malformations, with poor or absent lens differentiation and disorganized, diminutive retinal development after treatment with 9 ng of XlNLRR-6 MO, injected at the two-cell stage. oc, outer cornea; le, lens epithelium; lf, lens fiber cells; gc, retinal ganglion cell layer; bc, bipolar cell layer; pc, photoreceptor cell layer (note that at this stage the rod and cone photoreceptors have not yet differentiated); pe, retinal pigment epithelium. Scale bar = 400 μm in L (applies to A–L); 100 μm in R (applies to M–R).

Download figure to PowerPoint

Upon histological examination, these grossly observable defects were borne out as deficits in organization and differentiation of both the lens and retina (compare Fig. 6M with N–Q). In many cases, the lens was absent or abnormally small; those specimens possessing lenses displayed a multilayered, poorly differentiated anterior epithelium and few, poorly differentiated fiber cells (see Fig. 6N–O). Retinal defects included missing structures, particularly in the ventral region, and poor differentiation of the photoreceptor, bipolar, and retinal ganglion cell layers (see Fig. 6N–Q). Furthermore, in all XlNLRR-6 MO-treated specimens examined, the retinal pigment epithelium was either missing or incomplete. In addition to lens and retinal defects, knockdown of XlNLRR-6 yielded poor differentiation of the cornea; this normally thin, squamous bilayer of epithelium remained thick and cuboidal, similar to the surrounding head ectoderm.

Eye morphologies after MO treatment were classified into three categories: (1) specimens showing no gross eye defects (see Fig. 6B); (2) those with minor defects (see Fig. 6C,D) in diminished pigmentation (i.e., eye appears “pale” when compared with the contralateral, untreated eye), small eye size (i.e., eye is 40–70% of contralateral eye diameter), or minimal ventral dysgenesis (ventral retina is disorganized or absent); and (3) those with severely diminished, dysgenic, or missing eye structures (see Fig. 6E,F). Using these categories, sample sets were compared across dosages to examine the effect of increasing MO treatment versus the baseline defects observed in control animals and in those injected with control MO (see Fig. 7). Control MO treatments did not result in appreciable defects at the 9-ng or at the excessively high 45-ng dose with respect to uninjected control animals (Fig. 7A). At 4.5-ng, 9-ng, and 18-ng doses of the XlNLRR-6 MO, observed defects increased with each dose and were significantly more prevalent than any defects observed with doses of the control MO (Fig. 7B). These results suggest that the XlNLRR-6 MO has a specific effect, independent of general toxicity due to the introduction of MO or injection methods.

thumbnail image

Figure 7. Dose-dependent effects of Xenopus laevis neuronal leucine-rich repeat proteins (XlNLRR) -6 morpholino oligonucleotide (MO) on eye development. A: Development of the lens and retina by stage 35–38 in uninjected embryos and those injected with control MO. B: Effects of increasing doses of XlNLRR-6 MO, injected into one cell at the two-cell stage. C: Reduction of XlNLRR-6 MO-induced defects after co-injection with graded quantities of XlNLRR-6 MO mutant RNA (MM RNA). Error bars indicate standard error.

Download figure to PowerPoint

To ensure further the specificity of the XlNLRR-6 MO effects, an “MO mutant” (MM) XlNLRR-6 transcript was generated, in which six third-base substitutions were introduced into the second through seventh codons of XlNLRR-6 while preserving the amino acid sequence (see the Experimental Procedures section; Heasman,2002). Dose-dependent, base-substitution RNA rescue has been used as a stringent control in recent developmental studies as confirmation that MO knockdown effects are specific for the targeted transcript (Hashiguchi et al.,2004; Chung et al.,2005; Ando et al.,2005). Alternative controls for MO specificity have used antibody-based detection of reduced protein levels. A confound to using antibody-based controls for the XlNLRR-6 MO lies in the homogeneity of the proteins within the NLRR family. Amino acid sequence alignment of the various members of the protein family show an overwhelming similarity in many domains across multiple regions of each protein within the family, and contributed to the robustness of the phylogenetic analysis shown in Figure 3. However, because these proteins bear striking similarity to each other throughout most of their primary sequence, except at the extreme terminal ends, there is concern that an antibody-based assay for XlNLRR-6 would cross-react with related NLRR proteins. Fortunately, because the N-terminal of each protein is substantially different within approximately the first 50 amino acid residues, this similarity does not present a problem with the specificity of morpholino binding (as the MO only targets the first seven codons).

The XlNLRR-6 MM RNA transcript, to which the XlNLRR-6 MO is not likely to bind, was co-injected in graded quantities into two-cell embryos with 9 ng of XlNLRR-6 MO. By stage 35, these embryos had developed significantly fewer eye defects than those treated with a comparable dose of XlNLRR-6 MO alone, and as the dosage of MM RNA increased from 250–1,000 pg, the XlNLRR-6 MO-mediated defects were nearly eliminated (see Fig. 7C). Furthermore, the 500-pg dose of MM RNA was able to reduce demonstrably the severe defects caused by an 18-ng dose of XlNLRR-6 MO (compare Fig. 7B,C). This dose-dependent ability of XlNLRR-6 transcript to reduce the MO effect indicates the specificity of the XlNLRR-6 MO treatment.

XlNLRR-6 Is Required Independently by the Lens and Retina for Late Induction/Differentiation

Considering the spatiotemporal expression and MO defects associated with XlNLRR-6, it is clear that XlNLRR-6 is necessary for late induction of the lens and differentiation of cell layers in the retina. Previous studies have established that early and late induction of the lens depends on signaling from the retina (Henry and Grainger,1987, 1990), and some reciprocal requirement of the lens for proper retinal development has also been demonstrated (Hyer et al.,2003). Taken together, these studies provide three possible explanations for the role of XlNLRR-6 in lens and retina development: (1) XlNLRR-6 is required for retinal differentiation, and lens defects observed after knockdown are secondary due to the disruption of key signals from the retina; (2) XlNLRR-6 is required for late lens development, and secondary retinal defects ensue if knockdown disrupts the lens; and (3) XlNLRR-6 plays a role in both lens and retinal development. In this latter case, it would also be useful to establish the extent to which the role of XlNLRR-6 is autonomous in the lens versus the retina.

To test these hypotheses, reciprocal presumptive lens ectoderm (PLE) transplants were conducted (after Henry and Grainger,1987). Embryos were injected in one cell at the two-cell stage with either 9 ng of red lissamine-tagged XlNLRR-6 MO or 200 pg of rhodamine green dextran (as a differential lineage tracer) and allowed to develop to early neurula stage 14, the earliest stage at which XlNLRR-6 transcription is detectable and early lens induction is taking place. PLEs were traded between MO-treated and dextran-labeled embryos, generating animals whose cornea and lens would be derived from the donor tissue transplant, and whose optic cup and retina would develop from the host tissue (see Fig. 8A). These embryos allowed examination of the effect of XlNLRR-6 knockdown in the lens and retina independently of each other. To establish a baseline of developmental perturbation caused solely by the injections and surgery, reciprocal transplants were also conducted between animals injected with 9 ng of control MO and 200 pg of rhodamine green dextran at the two-cell stage.

thumbnail image

Figure 8. Reciprocal presumptive lens ectoderm (PLE) transplants. A: Schematic illustrating the surgical strategy. See the Experimental Procedures section for details. PLE transplants were conducted at neurula stage 14. B: Contralateral (i.e., unoperated) eye from a Xenopus laevis neuronal leucine-rich repeat proteins (XlNLRR) -6 morpholino oligonucleotide (MO) host/dextran-labeled donor PLE transplant specimen showing normal morphology. C: Operated eye from specimen shown in B, illustrating a dysgenic defect. D,E: Red and green (dextran) channel images of the specimen shown in C, showing the successful positioning of the dextran-labeled transplant at the position of the eye (arrowhead indicates green-labeled lens in E). F: Contralateral (i.e., unoperated) eye from a dextran-labeled host/XlNLRR-6 MO-treated donor PLE transplant specimen. G: Operated eye from the specimen shown in F. H,I: Red and green channel images of the specimen from G, showing the successful positioning of the red MO-labeled transplant at the position of the eye (arrowhead indicates red-labeled lens in H). J–O: Ten-micrometer sections of eyes resulting from transplant experiments at the same developmental stages shown in B–I. J: Normal retina from a control MO-treated host/dextran-labeled PLE donor transplant specimen. Proper differentiation of neuron and pigmented cell layers is indicated. K: Mild ventral defects in an XlNLRR-6 MO host/dextran-labeled PLE donor transplant specimen. Disorganized ventral neurons and absent ventral pigment epithelium are characteristic of this defect category (compare with J). L: Severe retinal defects in an XlNLRR-6 MO host/dextran-labeled PLE donor transplant specimen. Neural disorganization is evident throughout the retina, which is also diminished in size versus control (compare with J). This specimen's defect is also accompanied by lens disorganization. M: Normal lens from a control MO transplant specimen. The single-layer lens epithelium, equatorial margin of fiber cell formation, and differentiated fiber cell mass are indicated. N: Lens from dextran-labeled host/XlNLRR-6 MO-treated PLE donor specimen showing differentiation defect. Multilayered lens epithelium and fiber cell disorganization are indicated (compare with M). O: Whole eye section from a dextran-labeled host/XlNLRR-6 MO-treated PLE donor specimen showing lens and minor retinal differentiation defects (see text). bc, bipolar cell layer; co, cornea; dl, disorganized retinal cell layers; gc, ganglion cell layer; ip, inner plexiform layer; ld, lens defect in fiber cell differentiation; le, lens epithelium; lf, lens fiber cells; ma, margin of lens equatorial zone; me, abnormal multilayered lens epithelium; on, optic nerve; op, outer plexiform layer; pc, photoreceptor cell layer; pe, retinal pigment epithelium; vd, ventral disorganization. Scale bar = 400 μm in I (applies to B–I); 50 μm in O (applies to J–L,O); 20 μm in N (applies to M–N).

Download figure to PowerPoint

Figure 8B–E shows a sample MO host embryo at stage 35, with green-labeled ectoderm visible in the lens. Likewise, Figure 8F–I illustrate a reciprocal transplant specimen, with red-labeled MO donor ectoderm visible in the lens. Each transplant specimen was examined in this manner to ensure that the transplanted PLE tissue ended up in the correct position overlying the eye cup. At this stage, each transplant specimen was sectioned, stained with hematoxylin, and examined for lens and retina tissue morphology. Figure 8J illustrates a normal retina, with all of the appropriate cell layers indicated. Example defects observed in retinal development, particularly in the XlNLRR-6 MO host embryos, are shown in Figure 8K,L. Defects included ventral disorganization of neuron cell layers and pigmented epithelium (Fig. 8K), a more mild phenotype, and severe diminishment and disorganization of the entire retina (Fig. 8L; this specimen also shows a lens developmental defect). When compared with the normal morphology, these retinas do not have clearly demarcated inner and outer plexiform layers (which normally consist of predominantly cell processes and interneurons), preventing identification of individual neural cell layers. In XlNLRR-6 MO donor embryos, most of the lenses developed at a normal size, but most of them demonstrated defects in late differentiation. When compared with a normal lens (Fig. 8M), these lenses possess a multilayered lens epithelium and poorly differentiated fiber cells, which do not completely elongate and associate to form a proper fiber mass in the center of the lens (Fig. 8N). Nearly 40% of the XlNLRR-6 MO PLE donors also developed with retinal disorganization, a phenotype illustrated in Figure 8O (compare neural retinal cells with Fig. 8K).

Control and XlNLRR-6 MO PLE donor-derived lenses were examined for late lens fiber gamma-crystallin markers by means of immunohistochemistry at stage 35 (Henry and Grainger,1990). The control lens illustrates a well-defined primary fiber cell mass, developing secondary fiber cells in the marginal zones, and a thin anterior lens epithelium (Fig. 9A). Samples of the “late-differentiation-defective” lenses formed from XlNLRR-6 MO PLE donor tissues (Fig. 9B,C) confirm that the anterior epithelium consists of aberrant, multiple layers (unstained by crystallin antibody). The morphologically poorly differentiated fiber cell mass in each MO-treated lens did stain for gamma-crystallin, confirming that the XlNLRR-6 knockdown defect in the developing lens occurs very late and does not disrupt the pathways leading to crystallin protein expression.

thumbnail image

Figure 9. Immunohistochemical analysis of Xenopus laevis neuronal leucine-rich repeat proteins (XlNLRR) -6 morpholino oligonucleotide (MO) -treated donor presumptive lens ectoderm (PLE) -derived lenses for gamma-crystallin expression at stage 35. A: Control MO-treated lens illustrating normal morphology and gamma-crystallin expression, indicated by the brown stain (see the Experimental Procedures section). Note the single, thin anterior epithelium and prominent fiber cell masses. B,C: Lenses from XlNLRR-6 MO-treated PLE donor specimens transplanted into control host embryos showing late differentiation defects. Multilayered lens epithelium and fiber cell disorganization are indicated. Fiber cells, although morphologically deranged, illustrate expression of gamma-crystallin protein markers. co, cornea; ld, lens defect in fiber cell differentiation; le, lens epithelium; ma, margin of lens equatorial zone; me, abnormal multilayered lens epithelium; pf, primary lens fiber cells; sf, secondary lens fiber cells. Scale bar = 20 μm in C.

Download figure to PowerPoint

Results of the transplant experiments are summarized in Figure 10. XlNLRR-6 MO hosts had lenses resembling lenses from the control transplant population (compare Fig. 10A,B), but their retinas were defective to mild and severe extents in over 70% of cases (Fig. 10A). The reciprocal specimens, with XlNLRR-6 MO donor PLE, displayed the late lens differentiation phenotype shown in Figures 8N and 9B,C and also had modest retinal defects (illustrated in Figs. 8O, 10A, right). The baseline results with control MO-treated hosts and PLE donors, which represents a low level of defects caused by the surgery alone, are shown in Figure 10B.

thumbnail image

Figure 10. Reciprocal transplant results. A: Lens and retina phenotypes observed with Xenopus laevis neuronal leucine-rich repeat proteins (XlNLRR) -6 morpholino oligonucleotide (MO) host and presumptive lens ectoderm (PLE) donor specimens at stage 35–38. B: Baseline phenotypes observed with control MO host and PLE donor specimens. Error bars indicate standard error.

Download figure to PowerPoint

DISCUSSION

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

Neuronally Expressed LRR Proteins Comprise a Complex and Multifamilied Group

Based on the protein sequence comparisons in Table 2 and the phylogram in Figures 2 and 3, it appears that there is some incongruity in the naming conventions used for vertebrate neuronally expressed, LRR-containing proteins. The leucine-rich repeat neuronal (LRRN) proteins numbered 1, 2 and 3 and the leucine-rich repeat transmembrane neuronal (LRRTN) proteins numbered 3 and 4 demonstrate considerable similarity in the phylogram and represent one related group of LRR-containing proteins. Another group is likely to be the LRRN-6 group shown in Figure 2C, which shows some divergence between the A and D-type members. Finally, the neuronal leucine-rich repeat (NLRR) proteins, along with the LRRN-5 proteins, all group together. Because two of these groups contain members with different naming conventions, it can be confusing to understand their true relationships based on original nomenclature. For example, the HUGO Gene Nomenclature Committee (http://www.gene.ucl.ac.uk/nomenclature/) only lists human LRRN proteins, although the NCBI databases include human proteins from both the LRRN and NLRR groups. As is shown in Figure 3, the two groups of proteins are disparate and likely deserving of distinct nomenclature.

Another incongruity evident from Table 2 is that the mouse NLRR-4 sequence only tenuously matches the other NLRR family sequences against which it was compared. The 34–41% protein similarities observed were due mainly to sequence alignment within the LRR regions of the proteins, and no significant alignments were observed outside of this region. Because many LRR-containing protein families demonstrate significant similarity among their LRR, the protein originally characterized as mNLRR-4 does not appear to belong in the NLRR protein family proper and deserves further comparison to other groups within the LRR superfamily. These observations are also apparent from the phylogram in Figure 2A, which shows that the NLRR-4 proteins are a distant outgroup from the other NLRR family members.

Within the NLRR family, it is clear that the previously characterized zfNLRR protein is not orthologous to any other of the NLRR proteins described (see Table 2; Fig. 3). However, it does bear a strong protein sequence similarity to the full-length protein of Xenopus W006. Based on the sequence and phylogenetic data, these two proteins have been proposed to constitute the NLRR-6 group, and we are naming the Xenopus protein encoded by clone W006 as XlNLRR-6.

XlNLRR-6 Knockdown Disrupts Differentiation of Cells in the Lens and Retina

Xenopus lens induction begins during gastrulation, when a planar signal from presumptive neurectoderm and a vertical signal from subadjacent tissues, including underlying mesoderm and endoderm, produce a competent region of placodal ectoderm for lens formation (Henry and Grainger,1990; Schaefer et al.,1999). This “early” phase of lens induction continues to bias the PLE throughout neurulation stages 14–19. At the time of neural tube closure, the optic vesicle, the precursor of the retina, has evaginated from the neural tube to contact the lens-competent placodal ectoderm. Signals from the optic vesicle constitute the “late” phase of lens induction and specify a smaller region of competent head ectoderm for lens formation (Grainger,1996; Schaefer et al.,1999; Henry et al.,2002). At tail bud stage 26, the specified lens ectoderm is committed to form a lens placode. Lens placode cells proliferate to form a lens vesicle by stage 32, which ultimately detaches from the overlying ectoderm and comes to lie within the optic cup. Cells in the center of the vesicle elongate and condense into primary fiber cells, while the distal layer of cells remains as the lens epithelium, the source of lens “stem cells” (McDevitt and Brahma,1973). Epithelial cells in the transitional zone at the equator of the lens elongate and are pushed toward the center of the lens to form crystallin-packed fiber cells (see Fig. 8M; McDevitt and Brahma,1973).

During the stages of late lens induction, the optic vesicle further undergoes differentiation to form the eye cup and will ultimately give rise to retinal cell layers. The original spheroid shape of the optic vesicle invaginates from its distal hemisphere to form a concave, bilayered optic cup. The outermost layer of cells develops as the retinal pigment epithelium, while the innermost layer ultimately forms the three principal neural layers of the retina: the retinal ganglion cells, bipolar neurons, and photoreceptor cells (see Fig. 8J; reviewed by Chow and Lang,2001). As discussed above, the optic vesicle, optic cup, and retina are required for sustained induction of lens development and lens maintenance. Recent studies have shown that a reciprocal interaction from the lens ectoderm is required for proper morphogenesis of the optic cup and retina as well, in part by supplying retina-sustaining growth factors (Hyer et al.,1998, 2003).

In the retina, XlNLRR-6 is required for late differentiation, specifically in elaboration of neuron layers and the pigmented epithelium. The ventral retina, which differentiates later than the dorsal retina, is most sensitive to MO-mediated knockdown. XlNLRR-6 knockdown in the retina alone does not appear to affect lens development above baseline levels (Fig. 10). Therefore, while XlNLRR-6 knockdown disrupts differentiation of cells in the retina, it does not affect the ability of the eye cup to induce the lens.

In the lens, XlNLRR-6 is critical for late differentiation events in which equatorial cells destined to become fiber cells are unable to elongate and associate properly to form a crystallin mass as seen in control lenses (compare Fig. 8M with N,O). Additionally, XlNLRR-6 knockdown produces an abnormal, multilayered lens epithelium. Both of these lens defects may be due to a deficiency in the transition between stem-like, proliferating cells and those undergoing terminal differentiation (see below). XlNLRR-6 knockdown in the PLE/lens, while predominantly causing late differentiation phenotypes in the lens, also results in mild ventral developmental defects in the retina above baseline levels. These results suggest that inductive signals from the lens to the retina may be interrupted by XlNLRR-6 knockdown in the lens or that the lens disruption after knockdown occurs upstream of necessary lens-to-retina inductive signals. Therefore, while XlNLRR-6 plays an autonomous role in development of the lens, completely normal development of the retina also depends to some extent on XlNLRR-6 expression in the lens.

In addition to defects in the lens and retina, XlNLRR-6 knockdown also perturbs the differentiation of the cornea, as induced by the underlying optic cup and lens. In contrast to adjacent head ectoderm, the cornea at Xenopus stage 35 typically forms a thin squamous, bilayered epithelium that is optically transparent to allow for passage and focusing of light into the eye. Fibroblast-like keratocytes of neural crest origin subsequently migrate between the corneal epithelium and endothelium to secrete and maintain the collagen-rich corneal stroma (Hay,1977). XlNLRR-6 MO-treated embryos at stage 35 fail to develop the thin, specialized cornea epithelium, and instead possess a thick opaque, cuboidal epithelium overlying the eye similar to the adjacent head ectoderm (compare Fig. 6M with N–R). This result suggests that XlNLRR-6 is either intrinsically required for development of the cornea or that XlNLRR-6 knockdown in the lens and retina disrupts cornea-inducing signals from those structures. Because cornea defects were not prominent in the XlNLRR-6 MO PLE donor transplant specimens (not shown), it is likely that XlNLRR-6 knockdown in the eye affects induction of the cornea by means of the optic cup and is not intrinsically required in the cornea for proper development.

Eye Development Is Especially Sensitive to XlNLRR-6 Knockdown

Previous studies using MO-mediated knockdown of cell cycle control proteins and developmental signaling proteins in Xenopus have observed defects after doses equivalent to four to nine times more MO than were used in this study (Audic et al.,2001; Khokha et al.,2005). At the relatively modest 9- and 18-ng doses of XlNLRR-6 MO used, robust lens and retina development defects were observable. These results suggest that eye development is especially sensitive to levels of XlNLRR-6.

Related NLRR Proteins Have Been Implicated in Potentiation of EGF Signaling

Rat NLRR-3 has been shown to facilitate internalization of epidermal growth factor (EGF) ligand/receptor complexes and, thus, potentiates EGF signaling (Fukamachi et al.,2002). A nine-residue clathrin-mediated endocytosis motif in its intracellular domain (see Fig. 1A, white arrowhead) was shown to be necessary for NLRR-3 clustering and EGF-R internalization. XlNLRR-6 also possesses this endocytosis motif (see Table 1; Fig. 1).

Although no previous examination of NLRR proteins has been conducted in the lens or retina, it has been demonstrated that endogenous EGF-R is present in cells of the human lens epithelium and throughout developing retinal cells in the mouse and rat (Lillien,1995; Maidment et al.,2004; Rhee et al.,2004). EGF signaling has been implicated in cell proliferation and differentiation of lens epithelium into fiber cells and of retinal progenitor cells into neurons, glia, and photoreceptors (Lillien,1995; Maidment et al.,2004). Whereas the extracellular concentration of EGF in the developing vertebrate eye remains fairly low and constant, cells can alter the nature of their response to EGF signaling by modulating their expression of EGF-R and downstream signaling components (Lillien,1995; Maidment et al.,2004). By up- or down-regulating the magnitude of their responsiveness to EGF, progenitor cells in the eye may become biased toward sustained proliferation or differentiation.

The nature of the defects we observed in the lens and retina after XlNLRR-6 knockdown are consistent with failure in proliferation, in the cases of diminished lens and retinal mass, and are also consistent with failure of differentiation, such as in the late differentiation phenotype of lens fiber cells in the XlNLRR-6 MO PLE transplant cases shown in Figure 8N,O and Figure 9B,C, and the absence of organized, differentiated neural retina layers illustrated in Figure 6N–R. Further in vivo analysis could establish whether XlNLRR-6 interacts with known growth factor signaling pathways critical for vertebrate eye development.

XlNLRR-6 May Play a Critical Role in Brain Development

Whereas marked defects in the lens and retina after XlNLRR-6 MO-mediated knockdown were examined in this study, additional defects in forebrain development were also noted. Most cases with severe retinal differentiation defects showed concomitant disorganization of the normally polarized ventricular zone and some distal cortical cell layers (data not shown). Neurons in these layers did not develop the proper bipolar morphology, and the brain mass was diminished when compared with the contralateral (i.e., untreated) side of the brain. These results suggest that the roles of XlNLRR-6 in the forebrain and retina may be related in nerve morphogenesis. Future studies with the XlNLRR-6 MO should examine ventricular zone polarization, proliferation, and axon guidance in the forebrain to more fully understand the nature of the observed brain defects.

EXPERIMENTAL PROCEDURES

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

Animals

Adult Xenopus laevis were obtained from Xenopus I (Ann Arbor, MI) and NASCO (Fort Atkinson, WI). Fertilized eggs and embryos were prepared according to Henry and Grainger (1987). Embryos and larvae were reared as described in Henry and Mittleman (1995) and staged according to Nieuwkoop and Faber (1956).

Isolation of Full-Length XlNLRR-6

Two rounds of 5′-RACE (SMART RACE cDNA Amplification Kit, BD Biosciences Clontech, Palo Alto, CA) were used to amplify the full length of XlNLRR-6. Gene-specific, 5′-directed RACE and nested RACE primers (RACE primer, 5′-GCAATGCTGACTGGTCGACCTACCACTG; nested primer, 5′- GTCGACCTACCACTGACATAGAG; second fragment RACE primer, 5′-GTGCTTCCTTGGGCATCTCTG) were used as instructed by the kit to amplify successively longer fragments of XlNLRR-6 from total RNA collected by extraction from whole tissue by TriZol reagent (Invitrogen, Carlsbad, CA) from stage 32–36 hatchling larvae. The RACE fragment sequences were appended to the original W006 clone (Wolfe et al.,2004) using Sequencher software (Gene Codes, Ann Arbor, MI).

From the assembled RACE fragments, a full-length XlNLRR-6 first-strand cDNA was generated from stage 32–36 RNA using SuperScript II (Invitrogen). High-fidelity amplification was then conducted with Platinum Taq polymerase (Invitrogen) using the full-length reverse primer (5′-GATCCTGTCATCTGTAACGAG) and forward primer (5′-GTGGGAGAATCCATTCTGCAC), and the product was cloned into the pGEM-T Easy vector (Promega, Madison, WI).

Northern Hybridization Analysis

Total RNA was collected from stage 32–36 hatchling larvae by extraction from whole tissue using TriZol (Invitrogen), and poly-A RNA was purified using the Micro-polyA Purist Kit (Ambion, Austin, TX). Northern hybridization was conducted using glyoxylated RNA in a Bis-Tris-EDTA (BPTE) gel, followed by capillary transfer to nylon membrane and probing with 32P-labeled antisense XlNLRR-6 cDNA (as described by Sambrook and Russell,2001).

Analysis of NLRR Family Protein Sequences

Initial percentage similarities between protein sequences were determined using NCBI blastp and blast2seq programs (http://www.ncbi.nlm.nih.gov/). Forty-one selected protein sequences (see Fig. 2) were assembled and aligned with ClustalX software. The alignment was examined for areas of conservation among all of the selected proteins, and the N-terminal and C-terminal flanking regions (NTR, CTR), within which cysteine residues appear at highly conserved locations, were selected as alignable, nonrepeating domains for phylogenetic analysis.

Bayesian phylogenetic analysis was conducted using MrBayes software (http://www.mrbayes.net; Huelsenbeck et al.,2001). The 50,000 generations of Monte Carlo Markov Chain were run, and a consensus tree was assembled for the NTR/CTR alignment. After this initial analysis, 18 sequences corresponding to the NLRR Family orthology group and two neighboring outgroups were selected for analysis of full-length protein sequences (with short N-terminal and C-terminal exclusions based on highly variable, unalignable sequence) using the same parameters described above. A rooted consensus phylogram was constructed based on the NLRR full-length analysis.

In Situ Hybridization

Digoxigenin (DIG) -labeled RNAs were produced by PCR-amplifying XlNLRR-6 with standard primers for the SP6 and T7 promoter sites, followed by transcription with SP6 (antisense) or T7 (sense) RNA polymerase (Invitrogen or Roche, Indianapolis, IN). In situ hybridization was performed as previously described (Harland,1991), with the exception that 2 mg/ml glycine in phosphate buffered saline (PBS)/0.1% Tween was used to terminate Proteinase K digestion of specimens (Belo et al.,1997). Hybridized probe was detected by using anti–DIG-alkaline phosphatase (Roche) followed by the alkaline phosphate substrate BM-purple (Roche). After substrate development, pigmented specimens were refixed and photobleached in 1% H2O2, 5% formamide, and 0.5× SSC (Mayor et al.,1995).

Reverse Transcriptase-Polymerase Chain Reaction Analysis

Total RNA was extracted from groups of embryos at stages 1–3, 4–6, 10, 11.5, 14–15, 19–22, 25–27, 36–37, and 43–45 using TriZol reagent (Invitrogen). One microgram from each RNA pool served as a template for reverse transcription using SuperScript II (Invitrogen) and the XlNLRR-6 reverse primer (5′-GATCCTGTCATCTGTAACGAG). Polymerase chain reaction (PCR) amplification of cDNAs was then conducted using a gene-specific primer pair (forward, 5′-GTGGGAGAATCCATTCTGCAC; reverse, 5′-GCAATGCTGACTGGTCGACCTACCACTG) that yielded a 2.0-kb fragment of XlNLRR-6. Each PCR reaction was run on a 1% TAE agarose gel to establish the presence of XlNLRR-6 transcript in the original RNA pools.

MO Design

MOs are provided by Gene Tools, LLC (Philomath, OR). The control MO provided by Gene Tools (5′-CCTCTTACCTCAGTTACAATTTATA) is designed against a human globin intron and is not known to bind any other sequences. The anti–xNLRR-6 MO (5′-CCAGTCTGCTTTGTGTGGAGCATGG) was designed with assistance from Jon Moulton (Gene Tools, LLC) and the Gene Tools Web site (http://www.gene-tools.com/). This oligo is targeted to bases −2 to 23 of the xNLRR-6 transcript. Both MOs are tagged with 5′-lissamine, allowing red fluorescent tracking of tissues that have incorporated the MO.

Generation of MO Mutant mRNA

To demonstrate that the XlNLRR-6 MO specifically affects its transcript, RNA rescue experiments were undertaken. The MO binding site of the transcript was mutated to preserve the codon sequence of XlNLRR-6 but to avoid MO knockdown. A Xenopus codon usage chart (Nakamura et al.,2000) was used to establish appropriate codon mutations in the second through seventh codons of the transcript, while still ensuring appropriate tRNA coverage for those codons in the embryo. The original 25-bp sequence of XlNLRR-6, against which the MO is targeted, reads 5′-CCAUGCUCCACACAAAGCAGACUGG (the Met codon is underlined). After the start codon, the third bases of the next six codons were altered to read 5′-CCAUGCUGCAUACCAAACAAACAGG. Note that both sequences still code for amino acids MLHTKQT.

A forward PCR primer was designed to introduce these six-point mutations (5′-GCCCCACCATGCTGCATACCAAACAAACAGG) and was amplified using high-fidelity Platinum Taq (Invitrogen) and the full-length reverse primer (5′-GATCCTGTCATCTGTAACGAG). This sequence was cloned into the pCS2+ vector (Clontech) providing a 3′-polyadenylation signal, and transcribed using the SP6 MessageMachine kit (Ambion). Graded quantities of this mRNA were co-injected with graded quantities of XlNLRR-6 MO following the procedure described below.

Microinjection of Embryos

Two-cell embryos (stage 2 of Nieuwkoop and Faber,1956) were immersed in 1% Ficoll and fitted in wells in a dish of modeling clay. Graded quantities of MO, RNA, or rhodamine green dextran (10,000 molecular weight; Molecular Probes, Eugene, OR) were injected into one cell using glass microinjection needles with a Narishige micromanipulator (Narishige USA, East Meadow, NY) and Picospritzer II microinjector (General Valve Corp., Fairfield, NJ). After injection, embryos were allowed to recover in isotonic 1/20× NAM salt solution.

Reciprocal PLE Transplants

This procedure is as previously described (Henry and Grainger,1987). Embryos previously injected with either XlNLRR-6 MO, control MO, or rhodamine green dextran were allowed to recover in 1/20× NAM overnight until reaching beginning neural plate stage 14. These embryos were immersed in 3/4× NAM, demembranated, and fitted into clay wells. Using fine glass needles, the PLE was removed from the injected side of each embryo and the PLE from the reciprocal specimen was grafted in its place. Each reciprocal pair of specimens consisted of one MO-injected host with a dextran-labeled PLE donor and one dextran-labeled host with a MO-injected PLE donor. A flame-polished, siliconized fragment of glass coverslip was laid in place over each graft to ensure proper healing for 30–60 min. At 6 and 18 hr after transplantation, the solution was gradually changed with fresh 1/20× NAM.

Histological Preparation

Specimens were killed by extended treatment with MS222 (1:2,000 dilution; Sigma, St. Louis, MO), followed by fixation in MEMFA (3.7% formaldehyde, 100 mM MOPS, 2 mM EGTA, 1 mM MgSO4) or 4% paraformaldehyde and dehydration to 100% methanol. If not being used immediately, specimens in alcohol were stored at −20°C. Methanol was replaced with xylene, followed by a wash with 50% xylene/50% Paraplast Plus (Fisher Scientific, Pittsburgh, PA) at 58°C and two washes of fresh Paraplast. Each sample was oriented, anterior down, in Peel-A-Way paraffin embedding molds (Fisher), and the wax was allowed to harden. The specimens were sectioned in a microtome to 10 μm thickness; sections were mounted on gelatin-subbed glass microscope slides.

Wax was removed from the slides using xylene washes, followed by graded ethanol washes back to distilled water. Slides were immersed for 10 sec in Harris' hematoxylin (Fisher) and “blued” in running tap water and Scott's solution (after Humason,1972). After staining, graded dehydration washes were used to return the samples to xylene and coverslips were affixed using Permount medium (Sigma). Representative specimen sections shown in Figures 5, 6, 8, and 9 were selected as the central section through the lens and/or retina.

Immunohistochemistry

Sectioned specimens were immunostained for presence of differentiating lens gamma-crystallin proteins using polyclonal rabbit anti-lens antibody following the protocol of Henry and Grainger (1990). The antibody was generated in rabbit from soluble adult Xenopus lens proteins and was cross-absorbed with homogenized stage 19–24 embryonic tissues (in which crystallin proteins are not expressed) to ensure lens specificity; anti–gamma-crystallin specificity was confirmed by Western blotting, as described by Henry and Grainger (1990). Secondary detection used 1:5,000 donkey anti-rabbit IgG conjugated to horseradish peroxidase (GE - Amersham Biosciencees, Piscataway, NJ). Signal detection followed the protocol of the ImmunoPure Metal Enhanced DAB Substrate kit (Pierce, Rockford, IL). Completed immunostained slides were permanently mounted with Permount as described above.

Acknowledgements

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

The authors thank Max Telford at University College London and David Matus at the University of Hawaii for their assistance with multiple sequence alignment and Bayesian analysis of LRR proteins; Jon Moulton at Gene Tools, LLC, for assistance with designing the XlNLRR-6 MO; and Jeff Douglas, Jo Ann Cameron, Brian Walter, and Matthew Elkins at the University of Illinois for their advice during the preparation of this manuscript. A.D.W. received a Summer Student Fellowship from Fight For Sight, Inc.

REFERENCES

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