Many organisms display a remarkable ability to replace missing or damaged tissues (Stocum,2006). In some animals, parts of the eye, including the lens and retina, can be regenerated (Henry,2003; Henry et al.,2008; Tsonis,2008). In some species of newts and salamanders (Salamandridae), lenses can be regenerated via transdifferentiation of the dorsal iris pigmented epithelium, a process referred to as Wolffian lens regeneration (reviewed by Reyer,1954; Henry,2003). This process has also been described in one species of fish (Misgurnus anguillacaudatus, Sato,1961; Mitashov,1966), and possibly the chicken (van Deth,1940; McKeehan,1961; Niazi,1967; Wedlock and McCallion,1968; Okada,2000). Lens regeneration can also occur in the urodele Hynobius unnangso and members of the genus Xenopus through a process that has been described as “transdifferentiation” of the cornea epithelium (Ikeda,1936a,b,1939; Freeman,1963; Henry and Elkins,2001; Filoni et al.,2006). Whether this process represents a true process of transdifferentiation remains unclear (Henry,2003). Freeman (1963) divided the latter process into five stages in Xenopus, which resembles the process of embryonic lens formation. Once the original lens is removed, cells of the inner layer of the cornea epithelium become cuboidal (stage 1) and thicken as a placode (stage 2) to ultimately form a lens vesicle (stage 3) that differentiates primary (stage 4) and secondary fiber cells (stage 5) that contain lens crystallin proteins. This process is triggered by factors produced by the neural retina (reviewed by Henry,2003; Henry et al.,2008).
Though the cellular interactions involved in regulating lens regeneration are well characterized in Xenopus, relatively little is understood of the molecular events that underlie this process. Here we describe a study characterizing several hundred genes expressed over the first four days (stages 1–3) of the process of lens regeneration in the frog Xenopus laevis. This study builds on a preliminary report in which a smaller subset of genes were examined (Henry et al.,2002). In the previous study, we described some genes expressed during lens regeneration, but given the relative lack of bioinformatics data at that time, very few of these genes (45) could be related to other well-characterized genes on the basis of BLAST analyses. Therefore, we were not able to fully survey the biology and potential molecular mechanisms that underlie the process of lens regeneration in that system. Embryonic expression was examined for 15 of the genes described in that earlier study. This data set has been expanded considerably in the present study, and advances in sequence and genome analysis, particularly in the availability of expanded EST and genomic databases, have now allowed for the identification of orthologs and homologs characterized in a variety of metazoan systems. Using these tools, we have found that genes up-regulated during lens regeneration include a large number of developmental genes, such as various transcription factors, signal transduction pathway members, factors involved in RNA and protein processing and protein degradation, matrix metalloproteases (MMPs), cell cycle regulators, and others. Furthermore, we have examined the embryonic expression of most of these genes using automated, high-throughput in situ hybridization and show that a large number are also expressed during embryonic development in a variety of tissues, including the developing lens. These data support the argument that embryonic lens development and lens regeneration are related processes. Specific information pertaining to all of the genes characterized in this study is presented in a searchable, on-line database (www.life.illinois.edu/henry) available as a useful resource for the community of developmental biologists.
Understanding regeneration from an evolutionary perspective has been a difficult subject to address (Goss,1969,1992; Brockes et al.,2001; Sánchez Alvarado and Tsonis,2006; Brockes and Kumar,2008). Some have proposed that the widespread occurrence of regeneration amongst the Metazoa indicates that regeneration represents an ancient condition of metazoan biology (Brockes et al.,2001; Sánchez-Alvarado and Tsonis,2006; Brockes and Kumar,2008). To assess whether there may be conserved core molecular components of specific molecular processes underlying regenerative phenomena, we compared gene expression during lens regeneration in Xenopus with expression in other regenerating systems. These latter comparisons revealed a relatively small subset of genes with overlapping expression.
Genes Expressed During Lens Regeneration in Xenopus laevis
From a total of 840 clones selected for sequencing, 734 unique sequences were identified after all alignments were performed and contigs were assembled (see Supp. Fig. S1, which is available online, and www.life.illinois.edu/henry). In many cases, partial sequences could be extended further by comparisons with published Xenopus laevis ESTs. Sixty-four different mRNAs were represented multiple times in the group of clones selected for sequencing. Together, these sequences make up approximately one third of the genes represented in the library. The most prevalent species include clones B046 (7%), B097 (4%), B127 (2%), H097 (1%), and L127 (0.5%). The sequences represented by clones B025, B029, B118, C048, C076, H003, H030, H064, H109, J102, K128, and M202 are each represented in 0.4% of those analyzed. The other 47 unique sequences are each represented in 0.3% of the clones sequenced.
The combined collection of ESTs for the two species of Xenopus (laevis and tropicalis) is now ranked fourth (behind that of human, mouse and maize, on the organism list at NCBI dbEST (http://www.ncbi.nlm.nih.gov/dbEST/dbEST_summary.html), representing over 1.97 million ESTs (Showell and Conlon,2007; Carruthers and Stemple,2006). Likewise, a number of independent research groups also have made their own X. laevis sequence databases available (e.g., http://xenopus.nibb.ac.jp/). Using the extended X. laevis sequences, comparisons were made with those contained in NCBI's non-redundant database via BLASTn and BLASTx analyses. In addition, these sequences were compared with those contained in the X. tropicalis genome to search for orthologous or homologous genes. A high-quality draft of the X. tropicalis genome has been completed with >95% of expected genes present in the assembly (http://genome.jgi-psf.org/Xentr4/Xentr4.home.html) and an X. tropicalis genome report describing this project (Harland et al., in preparation, www.jgi.doe.gov/xenopus)) is now in preparation. Together, these resources permitted us to discern known and putative identities for the vast majority of the sequences contained in the data set. Exact or highly significant matches with characterized (65.4%) or uncharacterized (20%) proteins were found for a total of 627 sequences (see Fig. 1A). The other 107 sequences (14.6%) are completely novel with no significant matches in the databases. Specific details are available from a publicly accessible online database (www.life.illinois.edu/henry) and Table 1 lists some key representative genes. The genes listed in Table 1 represent more prevalent groups of genes that are grouped by process or function (e.g., transcription factors, cell signaling factors, MMPs, etc.), and are described further in the Discussion section. In an earlier study (Henry et al.,2002), only 11.6% of the 233 ESTs reported exhibited significant matches to other characterized proteins, another 27.5% displayed similarity to other uncharacterized ESTs, and 60.9% represented unknown transcripts.
Table 1. Representative Genes Expressed During Lens Regeneration in Xenopus laevis
Examples of genes expressed during Xenopus lens regeneration, within specific, prominent categories, as indicated. The clone IDs, NCBI accession numbers, and corresponding putative identities are listed. Genes listed in bold refer to those with embryonic expression shown in Figure 3. For a complete list of all genes recovered, including their sequences, similarities, and corresponding embryonic expression data, refer to www.life.illinois.edu/henry and Supplemental Figure S2.
Regulator of G-protein signaling 6
PHD finger protein 10
Pre-mRNA-processing factor 39
Eyes absent homolog 2
Heart- neural crest derivatives-expressed protein 1
NSFL1 cofactor p47
Mediator of RNA polymerase II transcription subunit 25
Iroquois related homeobox5, Irx5
Polymerase (RNA) II (DNA directed) polypeptide K
Enhancer of polycomb homolog 2
AN1-type zinc finger protein 3
CCAAT/enhancer-binding protein alpha
Homeodomain-interacting protein kinase 3
SET and MYND domain-containing protein 1
Adenosine deaminase domain-containing protein 1
Transcription initiation factor TFHD subunit II
Activated RNA polymcrase II transcription cofactor 4
Zinc finger protein 526
ETS domain-containing protein Elk-3
Zinc finger protein 420
Zinc finger CCHC domain-containing protein 4
Similar to PYM protein
Transcription termination factor, mitochondrial
Four and a half LIM domains protein 3
CCAAT/enhancer-binding protein alpha
Zinc finger protein 84
Prospero homeobox protein 1 (PROXI)
Wilms tumor protein 1-interacting protein
Homeobox protein EMX2
Forkhead box protein 11
LIM domain transcription factor 1, MO4
Zinc finger protein 479
Zinc finger MYND domain-containing protein 11
DNA replication licensing factor MCM4
Transcription initiation factor TFIID subunit 9
Homeodomain-interacting protein kinase 1
Transcription factor E2F5
Zinc finger CCCH domain-containing protein IIA
Homeobox protein EMX2
Regulator of G-protein signaling 6
Nuclear receptor coactivator 1
Rho-related BTB domain-containing protein 1
RAS guanyl-releasing protein 1
Insulin-like growth factor-binding protein 2
Protein patched homolog 1
snRNA-activating protein complex subunit 1
Rho GTPase-activating protein 10
Thyroid hormone receptor beta A
G-protein coupled receptor 84
26S protease regulatory subunit 8
Rho GDP-dissociation inhibitor 2
Myosin light chain kinase family member 4
Secreted frizzled-related protein 3
CKLE-like MARVEL transmembrane domain-containing protein 5
Cytokine receptor common subunit beta
Transforming growth factor beta regulator 1
TGF-beta receptor type III
Transforming growth factor-beta-induced protein ig-h3
RanBP-type & C3HC4-type zinc finger-containing protein 1
Ubiquitin specific protease 10
Ubiquitin specific protease 30
E3 ubiquitin-protein ligase MIB1
Adhesion-regulating molecule 1
UBX domain-containing protein 1
Sentrin-specific protease 8
E3 ubiquitin-protein ligase NEDD4
Proteasome subunit alpha type-4
Cold-inducible RNA-binding protein
T-cell surface glycoprotein CD3 delta chain
Interferon regulatory factor 2
Activator of 90-kDa heat shock protein ATPase homolog 1
Hypoxia-inducible factor 1 alpha
JNK1-associated membrane protein
Toll-like receptor 3
Tumor protein p53-inducible protein 11
Telomere length regulation protein TEL2 homo
Charged multivesicular body protein 2b
Chromodomain-helicase-DNA-binding protein 4
Probable global transcription activator SNF21.4
Vacuolar protein sorting-associated protein 72
INO80 complex subunit D
Kidney mitochondrial carrier protein 1
Creatine kinase B-type
Using the sequence information, we have linked specific Xenopus clones to corresponding Human UniProt IDs to recover Gene Ontology (GO) annotations for putative and known molecular functions, biological processes, and association with specific cellular components. Gene Ontology assignments were gathered using Human UniProt IDs that could be tied to 392 of the 734 unique sequences (www.uniprot.org; www.geneontology.org). Three hundred and seventy-six of those 392 genes possessed GO annotations. Figure 2 provides profiles of the genes represented in this set, which ranks the most abundant GO annotation assignments (a table listing all GO annotations for each sequence is included as Supp. Fig. S2).
Embryonic Expression Patterns
In a previous study, we found that a significant number of genes expressed during lens regeneration are also expressed during embryonic development (Schaefer et al.,1999; Carinato et al.,2000; Henry et al.,2002; Walter et al.,2004; Walter and Henry,2006; Wolfe and Henry,2006; Elkins and Henry,2006). In order to identify additional genes that might be expressed during development and more specifically in developing eye tissues, we carried out in situ hybridization analyses for most of the unique genes identified in this study. Overall, 703 unique genes were examined and 4.2% (31) of the 734 unique genes recovered did not yield DIG-labeled RNA probes (Fig. 1B). Of these 703 genes, 90% (634) exhibited some embryonic expression, while 10% (69) of the genes appeared to exhibit no detectable embryonic expression. Some representative expression patterns are shown in Figure 3, and all the expression patterns may be viewed at the online database (www.life.illinois.edu/henry). Supplemental Figure S1 also summarizes patterns of expression for each gene within certain tissues and regions. Four hundred and six genes (57.7%) exhibited some expression in eye tissues that included the retina (245, 34.9%) and/or the lens (329, 46.8%), with a very small subset of these exhibiting more highly restricted expression in lens tissue (six genes, 0.9%, see Fig. S1 and Fig. 3). Significantly, a number of the genes that exhibit lens-specific expression and those highly expressed in lens tissue represent novel, uncharacterized genes (refer to Supp. Fig. S1).
Comparisons with Patterns of Gene Expression in Other Regeneration Systems
Some investigators have proposed that the capacity to regenerate lost or damaged body parts represents an ancient conserved feature of metazoan biology (Brockes et al.,2001; Sánchez-Alverado and Tsonis,2006; Brockes and Kumar,2008). Though many extant animal phyla contain representatives that exhibit the ability to undergo various forms of regeneration, the extent to which these regenerative processes share certain conserved underlying molecular mechanisms is unclear. In fact, there have been a number of studies that have examined gene expression in regenerating tissues. In order to detect shared gene expression that might reveal conserved processes in different regenerating systems, we undertook a comparison of gene expression profiles using a number of published studies, restricting our analysis to those using Xenopus and other amphibians along with a few other species of fish and the planarian (see Table 2 and Supp. Fig. S3). The specific studies examined are listed in the Results section and in the legend for Supplemental Figure S3.
Table 2. Summary of Genes Expressed in Multiple Regenerating Systems
Genes expressed in multiple regeneration systems. The individual Xenopus lens regeneration clone IDs are shown to the far left. The known or putative identities based on similarity are shown to the far right. The genes are ranked according to frequency of occurrence in the various studies examined, which is also indicated in the table. These numbers include the present study. Comparisons were made to the studies of Sánchez Alvarado et al.,2002; King et al.,2003; Ishino et al.,2003; Padhi et al.,2004; Katogi et al.,2004; Tazaki et al.,2005; Reddien et al.,2005; Schnapp et al.,2005; Grow et al.,2006; Atkinson et al.,2006; Schebesta et al.,2006; Monaghan et al.,2007; Makarev et al.,2007; Pearl et al.,2008; Gorsic et al.,2008. The microarray data included from the publication of Grow et al. (2006) specifically includes the genes found to be up-regulated in their A, B, and C, sets of microarray comparisons. Genes listed in bold refer to those with embryonic expression shown here in Figure 3. For more specific details of these comparisons refer to the text and Supplemental Figure S3.
Cytochrome c oxidase subunit 1
Collagen alpha-2(I) chain
X.l. DNA ligase III isoform alpha
DNA ligase 3
Collagen alpha-1(I) chain
Cold-inducible RNA-binding protein
X.t. Tc1-like transposon
Proto-oncogene tyrosine-protein kinase Fyn
V-type proton ATPase catalytic subunit A
NADH dehydrogenase 1 alpha subcomplex subunit 1
Nucleolar protein 7
Coiled-coil domain-containing protein 58
40S ribosomal protein S3
Collagen alpha-1(V) chain
ER lumen protein retaining receptor 2
Transforming growth factor beta regulator 1
Transforming growth factor-beta-induced protein ig-h3
X.l. XFG 5-1 and XFG 5-2 genes for zinc finger proteins
Transcription factor E2F5
Dual specificity protein kinase TTK
Calcium/calmodulin-dependent protein kinase type II alpha chain
Nuclear cap-binding protein subunit 2
Sulfotransferase family cytosolic 2B member 1
X.l. XFG 5-1 and XFG 5-2 genes for zinc finger proteins
Uncharacterized protein C9orf85
Proteasome subunit alpha type-4
V(D)J recombination-activating protein 2
HEAT repeat-containing protein 2
Splicing factor 45
X.l. tbud02 mRNA
Leucine-rich repeat neuronal protein 1
X.l. gene for tudor repeat protein Xtr
Tetratricopeptide repeat protein 39B
Those genes that have significant homologs or orthologs with those recovered during Xenopus lens regeneration are indicated with colored boxes in Supplemental Figure S3, along with the corresponding accession numbers when available for each study examined. The genes are ranked according to frequency of expression amongst the various studies examined with the most prevalent genes listed at the top of Supplemental Figure S3. A summary list of all the genes that appeared in multiple regeneration studies is included here in Table 2.
Molecular Characterization of Lens Regeneration in Xenopus laevis
This study represents the most complete analysis of gene expression to date that is specific to the early processes involved in Xenopus lens regeneration (Henry,2003; Henry et al.,2008). The list of genes represents those expressed during the first 4 days following removal of the original lens, which ultimately culminates in lens differentiation and synthesis of lens crystallin proteins (up to Freeman stages 3–4 of lens regeneration; Freeman,1963; Brahma and McDevitt,1974; Henry and Mittleman,1995). The early stages of lens regeneration presumably involve processes of wound healing, cellular dedifferentiation, and the initial events associated with re-programming and re-differentiation of progenitor cells along the lens pathway.
GO annotation of the characterized genes provides general knowledge of the processes involved in lens regeneration (Fig. 2). Early events are dominated by expression of developmental genes that play roles in the regulation of transcription and in various cell signaling pathways (see Fig. 2 and Table 1). Several clusters of genes involved in specific processes appear to be prevalent amongst those identified, including: transcription factors and other genes involved in RNA synthesis and processing, numerous genes encoding integral membrane and transmembrane proteins, in particular components of prominent conserved cell signaling pathways, genes involved in RNA and protein processing, transport, and degradation including the ubiquitin/proteasome pathway, matrix metalloproteases (MMPs), regulators of the cell cycle, general metabolism, apoptosis, and genes involved in chromatin remodeling, inflammatory/stress responses, and immune responses.
Many of these genes are known to play important roles in eye/lens development. These genes include genes encoding gamma- (clone B025) and beta-crystallin proteins (clone M159), which are prominent in differentiated lens fiber cells (see Table 1 and Fig. 3). In addition, members of a number of signaling transduction pathways, transcription factors, and other genes known to be important for eye/lens development are also present (Kazlauskas,1994; de Iongh et al.,2001; Beebe et al.,2004; Lang,2004; Schlessinger,2004; Lupo et al.,2005; Thisse and Thisse,2005; Grogg et al.,2005,2006). Representatives are listed in Table 1 and described further below (see also Supp. Fig. S1 and www.life.illinois.edu/henry).
A large number of transcription factors are expressed during lens regeneration, which are listed in Table 1. These include a number of DNA-binding zinc finger proteins that have not been well characterized (e.g., clone C070 = PHD finger protein 10; J055 = AN1-type zinc finger protein 3 or testis-expressed sequence 27; K125 = Zinc finger protein 526; L001 = zinc finger protein 420; L017 = Zinc finger CCHC domain-containing protein 4; M101 = zinc finger protein 84; P070 = Zinc finger protein 479; P074 = Zinc finger MYND domain-containing protein 11 or Adenovirus 5 E1A-binding protein, Protein BS69; and R279 = Zinc finger CCCH domain-containing protein 11A). Many of these transcription factors have not been previously implicated in the process of lens cell determination or differentiation, but some factors, such as Prox1 (clone M103), have been shown to play key roles in lens development (Cvekl and Piatigorsky,1996; Wigle et al.,1999; Ring et al.,2000).
Cell Signaling Factors
Several elements of key signaling pathways are also represented in the genes expressed during lens regeneration, which are also recorded in Table 1. These include regulators of G-protein signal transduction, Rho and Ras signal transduction, Wnt, BMP, Retinoic acid, Hedgehog, TGF-beta, and FGF signal transduction, as well as others.
For instance, one clone (W005) encodes a protein similar to fibroblast growth factor receptor 3 (FGFR3), and another encodes sprouty homolog 2 (which may act as an FGF antagonist = clone K146). Fibroblast growth factors, known to play important roles in lens development, have also been shown to play important roles in the process of Wolffian lens regeneration in newts (i.e., FGF2, Kodama and Eguchi,1995; Hyuga, et al.,1993; Del-Rio Tsonis et al.,1997,1998; McDevitt et al.,1997) and in cornea-lens transdifferentiation in Xenopus (i.e., FGF1, Bosco et al.,1994,1997).
One clone represents a gene similar to bone morphogenetic protein 5 (BMP5 = N053), and another represents Sclerostin domain-containing protein 1 (K137), which acts as an inhibitor of BMP signaling and is also known to enhance Wnt signaling (Itasaki et al.,2003). Another clone represents a gene similar to TGF-beta receptor type III (K077), and clone K008 represents a gene encoding transforming growth factor-beta-induced protein ig-h3. Members of the TGF beta family (including specific bone morphogenetic proteins, such as BMP4 and BMP7) have been shown to play important roles in lens development and regeneration (Beebe et al.,2004; de Iongh et al.,2001; Grogg et al.,2005). Grogg et al. (2005) demonstrated that inhibition of the bone morphogenetic protein (BMP) pathway can also trigger lens formation in newt ventral iris pigmented epithelial cells.
Other ESTs represent Wnt7b (clone L087), and two different genes related to Secreted Frizzled-related protein 3 (clones J077 and L224), as well as one related to Secreted Frizzled-related protein 5 (clone P153), which act as modulators or antagonists of Wnt signaling. Wnt signaling (including specifically Wnt7b) has recently been shown to play important roles in lens development and crystallin expression (Graw,1996; Stump et al.,2003; Ang et al.,2004; Lovicu and McAvoy,2005; Chen et al.,2008).
In addition, the gene encoding retinoid X receptor (RXR) gamma (clone N013) was present along with a gene thought to interact with RXR and thyroid hormone receptors (clone H128 = 26S protease regulatory subunit 8), and a gene encoding Mediator of RNA polymerase II transcription subunit 25 (a subunit required for RNA pol II transcription that is required for RXR alpha and RAR alpha transcriptional regulation = clone H061). Retinoids, including retinoic acid (RA) and their receptors (RXRs and RARs), play key roles in the development of eye tissues including the retina, iris, and lens (Manns and Fritzsch,1991; Kastner et al.,1994; Graw,1996; Cvekl and Piatogorsky,1996; Li et al.,1997; Gopal-Srivastava et al.,1998; Enwright and Grainger,2000; Wagner et al.,2000; Kawamorita et al.,2002). Tsonis et al. (2000,2002) found that antagonists of retinoic acid receptors (e.g., RARα) and chemicals that inhibit the synthesis of retinoic acid (i.e., disulfiram) disrupt the process of Wolffian lens regeneration in the newt Notophthalmus viridescens, generally inhibiting this process, but sometimes leading to the formation of ectopic lenses. Interestingly, expression of Six-3 in the presence of retinoic acid triggers lens formation in ventral iris pigmented epithelium (Grogg et al.,2005), which otherwise cannot form a lens.
One clone represents the gene encoding the receptor for hedgehog protein, patched-1 (H035). Tsonis et al. (2004) have shown that both Sonic hedgehog (SHH) and Indian Hedgehog (IHH) and their receptors, patched-1 and patched-2 are expressed in both developing and regenerating lenses in the newt. Furthermore, disruption of hedgehog signaling perturbs Wolffian lens regeneration.
Furthermore, there are components of the MAPK/ERK signal transduction cascade, including: MAP kinase-interacting serine/threonine-protein kinase 2 or MAP kinase signal-integrating kinase 2 (Mnk2, clone R153) and MAPK activator 2 (clone R359). Other genes activate the MAPK/ERK pathway, including RAS guanyl-releasing protein (D032), and UPF0485 protein C1orf144 (alt name Putative MAPK-activating protein PM18/PM20/PM22; clone P099). It is well known that receptor tyrosine kinases such as EGF and FGF receptors signal via MAPK (Kazlauskas,1994; Schlessinger,2004; Thisse and Thisse,2005). Certain TGF-beta family members also signal via the MAPK cascade. Wang et al. (2008) demonstrated that MAPK signal transduction is essential for vitreous-induced lens fiber cell differentiation. Another clone encodes a gene involved in regulating Delta-Notch signaling (Mind bomb, clone M206).
A few novel elements of potential signal transduction pathways were also recovered, including a probable G protein–coupled receptor 84 related to the rhodopsin family of G protein–coupled receptors (clone H127), which is specifically expressed in developing lens cells (Fig. 3). In addition, there are a number of other transmembrane and integral membrane proteins (refer to Supp. Fig. S1).
The presence of genes involved in retinoic acid signal transduction, FGF signaling, Hedgehog, Wnt, MAPK, BMP, and TGF-beta signaling, which represent proteins known to play important roles in eye/lens development and regeneration in a variety of systems (Lang,2004; Lupo et al.,2005; Grogg et al.,2006), suggests that these signaling pathways may also play key roles in lens regeneration in Xenopus.
Matrix Metalloproteases (MMPs)
MMPs have been implicated in the process of regeneration in a number of systems (Carinato et al.,2000; Odelberg,2005) and a number of different MMPs are up-regulated during regeneration (Yang and Bryant,1994; Miyazaki et al.,1996; Yang et al.,1999; Kherif et al.,1999; Carinato et al.,2000; Kato et al.,2003; Vinarsky et al.,2005). Experimental studies suggest that MMPs play an important role in processes related to ECM remodeling, cell migration, and cell–cell interactions (Leontovich et al.,2000; Quinones et al.,2002; Vinarsky et al.,2005). MMPs are also expressed during the process of wound healing (Carinato et al.,2000; Berman1989; Matsubarar et al.,1991a,b; Fini et al.,1996,1998; Ye and Azar,1998; Gailit and Clark,1994; Barro et al.,1998; Salo et al.,1994). A large number of different matrix metalloproteases are expressed during lens regeneration (see Table 1), including: MMP-9 (clone B066), MMP-13 (clone B089), MMP-14 (clone L006), MMP-16, also referred to as MTMMP-3 (clone M125), MMP-18 (clone L080), and a gene similar to TIMP3-A (clone J103). Another gene, FACE-1 (clone N218), encodes a different zinc metalloprotease. In an earlier study, we examined the specific role of MMP9 (clone B066) during lens regeneration in Xenopus and found that this gene is not actually expressed in developing or regenerating lens tissues (Carinato et al.,2000), but rather is activated during corneal wound healing, which is associated with this process. Some of the other MMPs detected in the present study are also not expressed in embryonic lens tissues, including MMP-14 (clone L006, Harrison et al.,2004) and MMP-18 (clone L080, Tomlinson et al.,2008). On the other hand, MMP-13 (clone B089, Walter et al.,2004) and MMP-16 (clone M125, see Fig. 2) appear to be expressed in developing lens tissues and could play more direct roles in lens formation.
Protein Processing and Degradation
A large number of genes associated with protein processing and protein degradation were identified. In particular, a number of components of the ubiquitin/proteasome pathway are represented and are listed in Table 1. Expression of these genes may be associated with cellular remodeling and dedifferentiation, as the cornea epithelium prepares to re-differentiate along the lens pathway.
Stress, Inflammatory, and Immune Responses
A number of genes involved in stress, inflammation, and innate immune responses are also expressed in larval corneas undergoing lens regeneration. These genes are listed in Table 1. Expression of these genes may be related to the process of wound healing that accompanies the process of lens regeneration. On the other hand, these genes may play more direct roles in the formation of a new lens. Pearl et al. (2008) have shown that expression of genes associated with stress response pathways is correlated with successful limb regeneration in Xenopus; specifically, elevated Hsp60 and Hsp70 expression is associated with successful limb regeneration. It has further been shown that a missense mutation in Hsp60 in nbl (no blastema) mutant Zebrafish prevents successful fin regeneration (Makino et al.,2005). Hsp70 is also highly expressed during limb regeneration in the axolotl (Levesque et al.,2005) and the regenerating arms of starfish (Patruno et al.,2001).
Studies suggest that the process of inflammation can promote regeneration in other systems (Filbin,2006). For instance, factors secreted by macrophages appear to be involved in promoting axonal regeneration (Lu and Richardson,1991; Meyer-Franke, et al.1998; Cai, et al.2002; Filbin,2006). In fact, the process of inflammation associated with injury of the lens promotes axonal regeneration in the optic nerve (Leon et al.,2000). Related processes have been shown to be important during Wolffian lens regeneration in newts, including the activation of antigen-presenting cells that engulf the original lens, and the activation of thrombin from prothrombin, which appears to promote re-entry to the cell cycle (Imokawa and Brockes,2003; Imokawa et al.,2004; Kanao and Miyachi,2006; Goodwin and Brockes,2006; Brockes and Kumar,2008). On the other hand, some investigators have proposed that the development of immune specificity and systems that promote inflammation, tissue repair, or scar formation may have contributed to the loss of regenerative capacity in many vertebrate lineages (Brockes et al.,2001; Harty et al.,2003; Mescher and Neff,2004,2005,2007; Goodwin and Brockes,2006; Brockes and Kumar,2008).
Chromatin Remodelers/Transcription Regulators
A small number of genes recovered represent components involved in chromatin remodeling. These are all listed in Table 1. Chromatin remodeling complexes have been shown to play key roles in regulating gene expression important for various developmental processes (Deuring et al.,2000; Badenhorst et al.,2002; Domingos et al.,2002; Barak et al.,2003; Dirscherl et al.,2005) Therefore, it is possible that these factors may play key roles in regulating the process of lens regeneration.
Other Eye-Related Genes
Additional genes related to the eye were recovered, including clone H004, which encodes a gene similar to eyes absent homolog 2, and clone R147 encodes a gene similar to that encoding Retinitis pigmentosa 9 protein. In addition, there are genes known to play important roles in the cornea (e.g., Transforming growth factor-beta-induced protein ig-h3, clone K080). The latter encodes an adhesion protein, defects in which lead to corneal dystrophy (Skonier, et al.,1992; Escribano et al.,1994; Fujiki et al.,2001). Finally, clone P037 encodes N-acetylgalactosamine-6-sulfatase, and defects in the glycosaminoglycan metabolic process lead to corneal clouding (Tomatsu et al.,1991,2005).
Identification of Genes Expressed During Embryonic Lens Development
Significantly, a number of genes expressed during lens regeneration are also expressed during embryonic development, particularly in the developing tissues of the eye, which include the lens (Figs. 1B, 3, Supp. Fig. S1). These observations lend further support to arguments that the processes of embryonic lens development and lens regeneration are closely related (Schaefer et al.,1999; Mizuno et al.,1999,2005; Henry et al.,2002; Walter et al.,2004). A number of genes are expressed specifically in the developing lens or at very high levels in lens cells. Genes that appear to be expressed more specifically in the developing lens include those encoded by clones B025 (gamma crystallin B), H127 (probable G-protein coupled receptor 84), M159 (beta B-1 crystallin), P045 (novel protein), and R185 (novel protein). A number of other genes were found to be significantly up-regulated in developing lens tissues, including those represented by clones B105, B118, D043, E007, H186, J091, L118, M103, N216, P103, R378, W002, W006, W007, W008, W010, W012, as well as many others (see Fig. 3, and Supp. Fig. S1, and www.life.illinois.edu/henry). A number of the latter, as well as others, are highly expressed in the developing retina and in other specific tissues (refer to Fig. 3, and Supp. Fig. S1, and www.life.illinois.edu/henry). Further research is currently underway to understand the functions of the novel genes expressed within the developing lens.
Comparisons With Other Regeneration Systems
Regenerative processes are widespread amongst the Metazoa and authorities argue that regeneration represents an ancient feature of metazoan biology rather than a set of phenomena that arose independently through adaptive pressures in certain phylogenetic lineages (Goss,1969,1992; Sánchez-Alvarado,2000; Brockes et al.,2001; Sánchez-Alvarado and Tsonis,2006; Brockes and Kumar,2008). On the other hand, the range of processes by which tissues are replaced is rather diverse, involving those of cellular dedifferentiation, transdifferentiation, the differentiation of quiescent stem cells, as well as the deployment and proliferation of existing differentiated cells (Brockes et al.,2001; Sánchez-Alvarado and Tsonis,2006; Brockes and Kumar,2008). Other investigators have described general events and processes that appear to underlie regenerative phenomenon in different systems, including: wound healing, the initial formation of a specialized blastema or region of proliferating progenitor cells, as well as the creation or deployment of populations of multi-potent or uni-potent cells (e.g., stem cells or “un-differentiated” cells), as well as the proliferation of differentiated cells (Beck et al,2009; Brockes and Kumar,2006). These represent fundamental processes required for related aspects of tissue maintenance and survival. As there are a relatively small number of conserved developmental regulatory factors and signaling pathways that control key aspects of cell determination and differentiation in metazoans, it is likely that many of the same genes will be used to support regeneration in different systems (Sánchez-Alvarado and Tsonis,2006). Sánchez-Alvarado (2000) proposed that comparisons of gene expression might reveal these underlying conserved molecular mechanisms. In fact, only a few studies have attempted to compare gene expression between different regenerating systems (e.g., Wolfe et al.,2004; Tazaki et al.,2005; Beck et al.,2009, discussed further below).
In an attempt to discern shared components that may be important for regeneration, we have compared 21 different studies of regeneration examining three species of amphibians, two species of fish, as well as the planarian S. mediterranea. These examples encompass a variety of regenerating tissues, including several tissues that may be regenerated in Xenopus (e.g., those associated with the lens, limb, tail). The studies are each listed in Supplemental Figure S3 with the appropriate references to each also included here in the legend for Table 2. The comparisons did not reveal many similarities in terms of the suites of genes expressed in these different tissues, as compared with those expressed during lens regeneration in Xenopus. A total of 114 Xenopus genes expressed during lens regeneration were also found to be common to those expressed in these other regenerating systems. However, most of these (78) were only found to be expressed in one other regeneration study (Table 2 and Supp. Fig. S3). Twenty-four genes were found in two other studies, five genes in three other systems, six genes in four other studies, and a single gene was found in five other studies. The seven most prevalent genes included: cytochrome C oxidase subunit 1 (clone W010), glutamine synthase (clone C021), alpha-2(I) chain of collagen (clone C037), isoform alpha of DNA ligase III (clone C048), DNA ligase 3 (clone C077), alpha-1(I) chain of collagen (clone M202), and calmodulin (clone W009). The full list of these genes or their orthologs is found in Table 2 and Supplemental Figure S3.
The lack of any striking overlap between these studies may not be that surprising. There are several cautionary notes that must be considered in interpreting these types of comparisons. First, there is no consistent relationship in terms of the stages or ages of the regenerating tissues examined in these studies. In fact, it would be difficult to determine how to equate stages or phases of regeneration between studies using vastly different tissues and organisms. Second, different methods/platforms were used to collect these data, including microarrays, differential display, and subtractive hybridization, among others. Hence, the data sets are themselves limited and surveyed to different levels of completeness. Many of the sequences represented in these data sets, and thus the genes originally used for subsequent annotation in those studies, are not full-length and this presents an additional issue in that some similarities/redundancies may have gone undetected. This is further complicated by the fact that some authors did not release all of the relevant sequence data to permit full comparisons via BLASTn and BLASTx analyses conducted here. Despite these caveats, this represents the only available information describing these regenerating systems, and any prevalent, highly expressed genes could have been detected.
Clearly, more thorough analyses need to be conducted with a broad range of regenerating systems, preferably using whole genome microarrays, as these become available. Wolfe et al. (2004) compared genes expressed between lens and limb regeneration in Xenopus, and their screen of 16,000 clones representing genes expressed during Xenopus lens regeneration recovered only 13 common genes. Tazaki et al. (2005) compared genes expressed during Xenopus tail and limb regeneration and recovered only 15 genes. Beck et al. (2009) compared genes expressed during Xenopus limb regeneration (including the studies of Pearl et al.,2008; Tazaki et al.,2005; Grow et al.,2006). Fifteen genes are represented in common with those recovered by Tazaki et al. (2005) and 21 genes were represented in common with those up-regulated in the study by Grow et al. (2006). Interestingly, none of the latter genes are included in the set listed in Table 2.
It has been proposed that regenerative phenomena may simply represent an extension of developmental processes, perhaps related to the persistence of multi-potent developmental states present in certain larval or adult cells/tissues (Morgan,1901; Goss,1969,1992; Sánchez-Alvarado,2000; Brockes et al.,2001; Henry and Elkins,2001; Henry,2003; Sánchez-Alvarado and Tsonis,2006; Brockes and Kumar,2008). Hence, the degree of overlap may be similar to that occurring between the different developmental pathways that underlie the formation of the diverse structures comprising this comparative analysis.
Adult X. laevis were obtained from NASCO (Fort Atkinson, WI). Embryos were collected as previously described (Henry and Grainger,1987; Schaefer et al.,1999). Embryos stage 14 through stage 37 were fixed in MEMFA (0.1M MOPS, pH 7.4; 2 mM EGTA, pH 8.0; 1 mM MgSO4; 3.7% Formaldehyde), and stored in 100% methanol at −20°C prior to use for in situ hybridization (see below). All Xenopus developmental stages follow those of Nieuwkoop and Faber (1956).
Library Construction and Differential Colony Lift Hybridization
The construction of a subtracted cDNA library enriched for genes expressed during the process of lens regeneration in Xenopus laevis was described in Henry et al. (2002). The library includes genes up-regulated during the first 4 days of this process, which is triggered following removal of the original lens in stage-50–52 tadpole larvae. Biotin-labeled RNA, prepared from non-regenerating cornea tissue, was used as a driver against single-stranded cDNAs prepared from corneas undergoing lens regeneration to ultimately produce the subtracted cDNA plasmid library (prepared in pSport1, Invitrogen, Carlsbad, CA). The relatively high efficiency of the subtraction process is detailed in Henry et al. (2002). Samples of transformed DH12S E. coli containing plasmids of the un-amplified, subtracted cDNA library were spread on LB-ampicillin plates and individual colonies were picked at random for plasmid preparation. The insert size of all purified plasmids was checked via PCR with T7 and SP6 primers, followed by standard gel electrophoresis on 1% TAE agarose gels. Clones with a minimum size of 0.5 kb (up to 3.25 kb, with an average length of 1 kb) were selected for sequence analysis. The minimum size cut-off was chosen to ensure that ample sequence could be obtained for more reliable similarity analysis and to generate effective Digoxigenin (DIG)-labeled antisense RNA probes for in situ hybridization (see below). After an initial round of 93 clones were sequenced, it was found that three different sequences (accession numbers BM928880, BM92889, BM928898) were represented in multiple colonies, together accounting for nearly 13% of all clones represented in this sample. To avoid re-selecting these more abundant clones, 32P random prime–labeled probes generated from the above-mentioned sequences were used for differential colony lift hybridization. Probes were generated using the RTS RadPrime DNA labeling System (Invitrogen, Carlsbad, CA; see Henry et al.,2002). The mixed probe was then used to screen replicate colony lifts (MagnaGraph, Osmonics, Inc., Minnetonka, MN; Church and Gilbert,1984) and additional colonies were selected for sequencing only if their plasmid DNA did not hybridize to the mixed probe.
DNA Sequencing and GeneOntology Assessment
Sequencing was completed at the University of Illinois Biotechnology Center (Urbana-Champaign, IL). The universal T7 primer was used to obtain sequence from the 5′ end of each directional clone. Sequencher (ABI Prism, Foster City, CA) was used to analyze and edit the sequences. All sequences have been deposited in the NCBI database. All sequences were compared with Xenopus laevis ESTs contained in the NCBI database as well as those found in other Xenopus laevis databases (e.g., http://xenopus.nibb.ac.jp/) to assemble larger contigs (original and extended FASTA files may be downloaded at www.life.illinois.edu/henry). The extended FASTA file was then used for comparisons to all sequences in the NCBI database using BLASTn and BLASTx search programs (www.ncbi.nlm.gov, Altschul et al.,1997). In addition, sequences were also compared to the X. tropicalis genome (http://genome.jgi-psf.org/Xentr4/Xentr4.home.html). UniGene IDs were obtained for most of the clones in the data set on the basis of BLAST searches described above. We used UniGene (www.ncbi.nlm.nih.gov/unigene) to establish a standardized list of equivalent human UniProt IDs (www.uniprot.org). Gene Ontology (GO) assignments were subsequently made via http://www.geneontology.org/.
DIG-Labeled Probe Preparation
Digoxigenin (DIG)-labeled antisense and sense RNA probes were synthesized from the purified pSPORT1 plasmid DNA. Probes were generated using SP6 (antisense) and T7 (sense) RNA polymerase (Roche, Indianapolis, IN) from an initial PCR amplification of the cloned sequences, prepared using standard T7 and Sp6 primers. The probes were purified using NucAway spin columns (Ambion, Austin, TX). Probes that were 1 kb or greater in length were hydrolyzed as follows. One microgram of DIG-labeled RNA was added to an equal volume of DEPC-treated water. Two volumes of carbonate buffer (60 mM Na2CO3; 40 mM NaHCO3; pH 10.2) were added to the water-diluted DIG-labeled RNA and the probes were incubated for 20 to 25 min at 60°C. Hydrolysis was stopped by addition of 3M sodium acetate and 1% v/v acetic acid (ph 6.0). Finally, NucAway columns were used to purify the probes following the manufacturer's instructions (Ambion, Austin, TX).
In Situ Hybridization
In situ hybridization was performed using Xenopus laevis albino embryos (stages 14 to 37) in a Biolane HTI robot (Hölle and Hüttner, Tubingen, Germany). Groups of embryos were held in small 53 μm mesh bottom baskets fit into plastic trays placed in the robot, which controlled washing, mixing, and temperature. Prior to each run, the mesh baskets, trays, and robotic tubing were thoroughly cleaned using RNAse wash buffer (0.1N NaOH; 1 mM EDTA in DEPC-treated water). Following these washes, the trays and baskets were removed and rinsed twice with DEPC-treated water.
A few minor modifications were made to the general in situ protocol described by Harland (1991). Hybridization and washes were completed in the robot over a 2-day period with gentle agitation at room temperature, unless otherwise specified. The first day started by rehydrating embryos through methanol/PTw (1× PBS, pH=7.4, 0.1% Tween) gradient washes, followed by three PTw washes. Proteinase K (10 μg/mL) digestion for 15 min was followed by a glycine (2 mg/mL) wash and two PTw washes. Embryos were then fixed in 4% paraformaldehyde in deionized water (PFA), washed twice in PTw, and incubated in hybridization buffer (50% deionized formamide; 5× SSC; 1 mg/mL Torula RNA; 100 μg/mL Heparin; 1× Denhart's; 0.1% Tween 20; 0.1% CHAPS; 5 mM EDTA) at 60°C for 4 hr.
DIG-labeled probes were added by transferring baskets to 24-well cell culture plates and the individual probes were added to each well, followed by overnight incubation at 60°C. After draining and transferring the baskets back to the robotic trays on the second day, washes continued according to Harland (1991). Anti-DIG Fab antibody (Roche, Indianapolis, IN, 1:2,000 dilution) was added to the baskets for 2 hr at 30°C, followed by two 20-min washes and three 10-min washes in MAB (100 mM maleic acid; 150 mM NaCl, pH 7.5). Embryos were washed twice in AP buffer (100 mM Tris, pH 9.5; 50 mM MgCl2; 0.1% Tween 20; 5 mM Lavamisol) and finally transferred from baskets to 24-well cell culture plates. mRNA localization was detected by application of BM-purple substrate (Roche, Indianapolis, IN) and color development was done outside of the Biolane robot, which was carried out for variable lengths of time at room temperature or 4°C, depending on the individual samples. Color development was stopped by washing the samples in PBS and re-fixing them in PFA for 1 hr at room temperature. Control negative (sense) and positive (anti-sense) samples (typically clone B025) were included in each run.
Whole mount specimens were photographed using an Olympus SZX12 dissecting microscope and a Nikon CoolPix 995 camera. Color images were processed using Adobe Photoshop (Mountain View, CA).
The on-line database (www.life.illinois.edu/henry) is maintained by the University of Illinois-Urbana Champaign, and was originally created using Filemaker Pro (Filemaker, Inc, Santa Clara, CA). The information was subsequently exported for Web access using Ruby on Rails (http://rubyonrails.org). This is a publicly accessible, searchable database with all sequence data, similarities, and additional links (e.g., NCBI X. laevis accessions; JGI-related X. tropicalis genomic information, embryonic expression) related to each EST characterized in this study. In addition, one can download complete FASTA files of all EST sequences from the above web site.
Blast Comparisons of Genes Expressed in Various Regenerating Systems
BLASTn and BLASTx analyses were undertaken to compare the suite of genes expressed during lens regeneration in Xenopus with those expressed in other systems (Altschul et al., 1979). Given the large range of studies that have been carried out, we restricted our analysis to those using Xenopus (King et al.,2003; Ishino et al.,2003; Tazaki et al.,2005; Grow et al.,2006; Pearl et al.,2008), other amphibians (axolotl: Schnapp et al.,2005; Gorsic et al.,2008; newt: Atkinson et al.,2006; Makarev et al.,2007), as well as fish (zebrafish: Padhi et al.,2004; Schebesta et al.,2006; Monaghan et al.,2007; medaka: Katogi et al.,2004). We also included studies examining regeneration in the planarian (Schmidtea mediterranea) for which an extensive characterization has been undertaken (Sánchez-Alvarado et al.,2002; Reddien et al.,2005). These studies were undertaken in different species and examined different tissues at various stages of regeneration. Given the large number of permutations, the genes expressed during lens regeneration in Xenopus served as the baseline for all comparisons. Furthermore, given the manner in which the subtracted cDNA library was constructed to enrich for genes up-regulated during the process of Xenopus lens regeneration, these comparisons were only made against those populations of genes that were found to be up-regulated during these other regenerative phenomena or those found to be critical for normal regeneration as determined in functional assays. For comparisons made with studies using Xenopus laevis (e.g., King et al.,2003; Ishino et al.,2003; Tazaki et al.,2005; Grow et al.,2006; Pearl et al.,2008 for limb and tail regeneration), only significant matches with expect values <1 ×10-100 were accepted. This cutoff was chosen as it would include all exact matches (e.g., E = 0) as well as those that may have had issues related to misalignments due to sequencing errors, or would include genes that are very closely related. For those comparisons with studies examining other species (e.g., Schnapp et al.,2005; Schebesta et al.,2006; Roddy et al.,2008; Padhi, et al.,2004; Monaghan et al.,2007; Lien et al.,2006; Katogi et al.,2004; Gorsic, et al.,2008; Atkinson et al.,2006) significant matches were accepted with expect values ranging as high as 1 × 10-20 that extended over spans of at least 50 amino acids, and in those cases where the sequences had already been identified (annotated) with homologous gene names.
The authors thank Adam Wolfe, Thom Uebele, and Jeff Haas for their assistance in setting up the online database. This research was supported by NIH-NEI Grant EY09844 to J.J.H.