SEARCH

SEARCH BY CITATION

Keywords:

  • desmoplakin;
  • kazrin;
  • cell-adhesion;
  • nuclear matrix;
  • desmosome;
  • adherens junction;
  • plakins;
  • fertilization;
  • embryogenesis

Abstract

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

The cell uses the cytoskeleton in virtually every aspect of cell survival and function. One primary function of the cytoskeleton is to connect to and stabilize intercellular junctions. To accomplish this task, microtubules, actin filaments, and intermediate filaments utilize cytolinker proteins, which physically bind the cytoskeletal filament to the core proteins of the adhesion junction. The plakin family of linker proteins have been in the spotlight recently as critical components for embryo survival and, when mutated, the cause of diseases such as muscular dystrophy and cardiomyopathies. Here, we reveal the dynamics of a recently discovered plakin binding protein, kazrin (kaz), during early mouse development. Kaz was originally found in adult tissues, primarily epidermis, linking periplakin to the plasma membrane and colocalizing with desmoplakin in desmosomes. Using reverse transcriptase-polymerase chain reaction, Western blots, and confocal microscopy, we found kaz in unfertilized eggs associated with the spindle apparatus and cytoskeletal sheets. As quickly as 5 min after egg activation, kaz relocates to a diffuse peri-spindle position, followed 20–30 min later by clear localization to the presumptive cytokinetic ring. Before the blastocyst stage of development, kaz associates with the nuclear matrix in a cell cycle-dependent manner, and also associates with the cytoplasmic actin cytoskeleton. After blastocyst formation, kaz localization and potential function(s) become highly complex as it is found associating with cell–cell junctions, the cytoskeleton, and nucleus. Postimplantation stages of development reveal that kaz retains a multifunctional, tissue-specific role as it is detected at diverse locations in various embryonic tissue types. Developmental Dynamics 234:201–214, 2005. © 2005 Wiley-Liss, Inc.


INTRODUCTION

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

The cytoskeleton is highly interactive in virtually all aspects of cell function and survival. As a result, the cell has assembled an abundant array of transcriptional, translational, and posttranslational mechanisms to carefully control cytoskeletal morphology and function. One of the more recent findings encompassing cytoskeletal regulation has been the discovery of crosslinker proteins that act by crosslinking cytoskeletal filament systems (i.e., actin, microtubules, and intermediate filaments), linking cytoskeletal filament systems to intercellular junctions (i.e., adherens junctions, desmosomes, gap junctions, and tight junctions), and linking different types of cellular junctions to each other (Sobel, 1990; Yang et al., 1996, 1999; Karakesisoglou et al., 2000; Leung et al., 2002; Weber and Bement, 2002; Giepmans, 2004; Jang et al., 2005; Winder and Ayscough, 2005).

These intimate cross-cytoskeletal/junctional connections have been shown to be crucial in regulating the cell cycle, tissue development, wound healing, and overall tissue integrity (McLean et al., 1996; Boudreau and Bissell, 1998; Gallicano, 2001; Gallicano et al., 2001; Leung et al., 2002; Grossmann et al., 2004; Setzer et al., 2004; Weber et al., 2004; Nishizawa et al., 2005). One of the oldest known, and yet still growing, family of cytoskeletal crosslinking proteins is the plakin family (Leung et al., 2002). Seven plakin family members have been identified. They all share the ability to link cytoskeletal filaments to an intercellular junction complex, whereas a few members recently have been shown to link intermediate filaments to actin and tubulin (Yang et al., 1996, 1999; Karakesisoglou et al., 2000; Weber and Bement, 2002).

Plakins rely on distinct anchoring proteins such as those of the armadillo, integrin, or cadherin families to regulate their linkage to junctional complexes (Perez-Moreno et al., 2003; Setzer et al., 2004; Wang and Zhang, 2005). Recently, convincing evidence revealed a novel noncanonical plakin interacting protein, kazrin (kaz), which acts as a cytolinker protein linking periplakin to the plasma membrane in keratinocytes (Groot et al., 2004). Of interest, images in Groot et al. (2004) showed evidence that kaz brings desmosomal plakins such as periplakin and desmoplakin (DP) into close proximity with actin filaments near adherens junctions implying that kaz may act as a cytolinker functionally connecting adherens junctions and desmosomes (Groot et al., 2004).

Kaz was identified with a yeast two-hybrid screen using a small portion of periplakin as prey and a human keratinocyte library as bait. In humans, kaz was found to have four splice variants, termed kaz A–D. All are capable of binding plakins, and all contain a nuclear localizing signal (NLS), which is functional. In mice, two variants have been found in mouse expressed sequence tag databases to date, kaz A and B (Groot et al., 2004; personal searches). Since its initial discovery in keratinocytes, kaz has been found in numerous different tissue types (Groot et al., 2004). Consequently, the dual cellular localization of kaz (junctional and nuclear) and its broad spectrum of expression suggested that kaz may have multiple functions across many tissues. However, it was the colocalization of kazrin with DP that led us to generate a novel hypothesis; that is, kaz, like DP, may be involved in the early developmental process.

DP has been studied recently in great detail using homologous recombination to knockout gene function, both tissue specifically and globally. DP has also been knocked down in endothelial cells using siRNA (Gallicano et al., 1998, 2001; Vasioukhin et al., 2001; Zhou et al., 2004). Each of those investigations clearly demonstrated the importance of DP, especially in the three-dimensional formation of developing epidermis, neurectoderm, heart, and capillaries. Kaz may also be involved in some or all of these processes by linking desmosomes through DP to other junctional components. Periplakin has not been identified at early stages of development (Aho et al., 2004), leaving DP as the sole plakin that may bind kaz.

To begin testing our hypothesis, we subjected eggs, preimplantation embryos, and postimplantation embryos to a myriad of molecular, biochemical, and microscopic analyses. We not only show kaz in unfertilized eggs, we also show that it is a highly dynamic protein after egg activation and throughout development.

RESULTS

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

Kazrin in Mouse Eggs

To begin determining a function for kaz during development, we first determined its expression pattern in mouse eggs and embryos at specific stages of embryogenesis using reverse transcriptase-polymerase chain reaction (RT-PCR), confocal microscopy, and Western blot. Confocal microscopy with anti-kaz antibody revealed the presence of kaz within eggs arrested at meiotic metaphase II (Fig. 1). Specifically and unexpectedly, it was localized on the spindle apparatus in unextracted and detergent extracted eggs (Fig. 1A–F show extracted egg). Double labeling with antibodies to tubulin resulted in clear colocalization between kaz and spindle microtubules. Figure 1E,F quantifies the colocalization between tubulin and kaz. Pixel colocalization was analyzed and quantified by comparing four areas, all of exact size of 695.0 pixels, from three regions of the egg (Fig. 1E). Region 1 compared colocalization of kaz and tubulin within cytoplasmic regions inside the egg (nonspindle), whereas region 2 and 3, respectively, analyzed kaz and tubulin colocalization in the chromosome region (metaphase plate represented by 4′,6-diamidine-2-phenylidole-dihydrochloride [DAPI] staining) and kaz/tubulin colocalization on the spindle apparatus. Of the observed area within the egg, virtually no kaz/tubulin signal overlap was detected within the egg cytoplasm (0.9 ± 0.5 overlapping pixels) and a low kaz/tubulin colocalizing signal was found within the central region of the metaphase plate chromosomes (3 ± 1.5 overlapping pixels). On the spindle, however, the signal overlap was dramatically and significantly higher at 342 ± 42 overlapping pixels. Detergent extraction of eggs (shown in Fig. 1A–F) resulted in little difference in these overlapping signals when compared with eggs left intact, demonstrating that kaz is directly associated with the detergent-resistant cytoskeleton.

thumbnail image

Figure 1. Kaz colocalizes to the spindle apparatus. A–C: Low (A,B,C) and representative high magnification views (A′,B′,C′) revealed the cytoskeletal nature of kaz colocalization with the spindle in the detergent-extracted unfertilized egg (arrows). D: randomized areas of equal size were analyzed for pixel intensity and direct colocalization of antibody staining. E: Quantification of direct colocalization between tubulin and kaz. The spindle apparatus showed 100× more colocalized pixels than other areas within the egg. F: Pixel colocalization quantification of kaz and tubulin with DNA as measured by 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining revealed 2× more tubulin colocalized with DNA (130.6 ± 15 pixels, primarily at the ends of tubulin filaments) than kaz (63.2 ± 5 overlapping pixels). Scale bars = 10 μm in A (applies to A–C), 2.5 μm in A′–C′, 5 μm for G.

Download figure to PowerPoint

At high magnification, some kaz staining appeared directly on chromosomes aligned on the metaphase plate; however, when compared with tubulin/DAPI colocalization (130 ± 23 colocalized pixels), kaz/DAPI colocalization (63 ± 14 colocalized pixels) was approximately twofold lower (Fig. 1F). Moreover, when the DAPI signal was removed electronically, it was evident that some of the kaz/DAPI overlap could be attributed to kaz aligned on adjacent microtubules bound to DNA and not direct kaz/DNA alignment (data not shown).

We also observed kaz throughout the cytoplasm of eggs (Fig. 1A–C; nonspindle areas). The signal intensity for kaz deep within the cytoplasm appreciably decreased upon detergent extraction (nondetergent extracted egg not shown); however, a distinct signal intensity was detectable in extracted eggs, suggesting that kaz was not only associated with the tubulin cytoskeletal spindle but also internal cytoskeletal components, possibly with the cytoskeletal sheets shown to exist in vast quantities in eggs and early embryos (Capco and McGaughey, 1986; McGaughey and Capco, 1989; Gallicano et al., 1991, 1992; Gallicano, 2001). Cytoskeletal sheets have been shown to be composed of intermediate filaments covered by a yet-to-be-identified particulate material (Capco et al., 1993). Consequently, the association of kaz with cytoskeletal sheets is of high interest and will have to be pursued using a number of assays, including immunoelectron microscopy and co-immunoprecipitation.

Further confirmation of kaz in eggs was revealed by Western blot (Fig. 2). A strong band was observed at 47 kDa, which is the correct molecular weight for Kaz A and B. A second, less intense band was detected at 37 kDa, representing kaz C. Fractionation of eggs using a nonionic detergent to generate a detergent soluble fraction and a detergent-resistant cytoskeletal fraction, followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis revealed that levels of kaz were higher in the detergent-resistant cytoskeletal fraction (Fig. 2). The Western blot also revealed a faint doublet, suggesting that kaz A and possibly B were present; however, because of the low amount of protein from 125 eggs, it was difficult to consistently resolve the doublet of kaz A and B. Of interest, the Western blot revealed a band at the correct molecular weight for kaz C. Previously, only kaz A and B transcripts had been identified in mice (Groot et al., 2004). To further confirm our Western results, we subjected eggs to RT-PCR using primers to each kaz A–D as well as primers that recognize all isoforms. Kaz A was detected; however, other kazs were not. Message for kazs B and C may be present during oocyte maturation but not in the egg. We are currently exploring kaz expression and function at various stages of oocyte development.

thumbnail image

Figure 2. Kaz was analyzed in 125 mouse eggs by Western blot and 25 eggs for reverse transcriptase-polymerase chain reaction (RT-PCR). A: Eggs were separated into a detergent-soluble (sol) and detergent-resistant cytoskeletal fraction (CSK). Samples were subjected to sodium dodecyl sulfate, blotted onto polyvinyl difluoride membrane and probed for kaz using the anti-kaz antibody supplied by Dr. Fiona Watt. A doublet was detectable at 47 kDa, representing kaz A/B (Sol lane, asterisks), whereas a faint band at 37 kDa, representing kaz C, also was detected. B: RT-PCR using primers against kaz detected only one kaz band, representing kaz A. Our positive control revealed kaz A and B in mouse embryonic fibroblasts but not kaz C/D. RT-PCR bands were the exact pair sizes for kaz.

Download figure to PowerPoint

Kazrin Dynamics During Egg Activation

Using confocal microscopy, we found dramatic changes in kaz localization during egg activation (Figs. 3–5). As soon as 5–10 min after activation, kaz staining on the spindle was more diffuse when compared with unactivated eggs. By 10 min after activation, the spindle was virtually devoid of kaz staining, while juxtaposed, peri-spindle localization of kaz appeared to increase (Fig. 3A,B). A distinct enrichment of kaz staining was repeatedly and clearly apparent at the future site of the cytokinetic ring by 20 min after activation (Fig. 3C,D). This strip of kaz staining was found to extend perpendicularly ∼1–5 μm from the anaphase spindle (Fig. 3D). By 30 min, kaz staining distinctly localized to the area forming the cytokinetic ring (Fig. 4). To determine whether kaz was directly associated with the cytoskeleton during cytokinesis, kaz staining was analyzed in activated eggs that had been extracted with a nonionic detergent. Although the perpendicular strip of kaz decreased in size in extracted eggs, distinct staining persisted at the imminent site of the cytokinetic ring, suggesting that kaz is associated with the cytoskeletal components of cytokinesis during second polar body emission (Fig. 4C,D). By 45 min, as second polar body emission neared completion, a clear ring of kaz staining was evident on the second polar body. Consequently, these data offer support to a novel hypothesis: kaz is necessary for cytokinesis (Fig. 5A).

thumbnail image

Figure 3. Kaz localization changes after egg activation. A,B: Ten minutes after activation, the spindle is virtually devoid of kaz, whereas the perispindle area shows an increase in kaz staining (arrows). Anti-tubulin and 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining clearly show the metaphase/anaphase transition. C,D: By 20 min after activation, the perispindle staining extends perpendicularly from the central region of the spindle where the future cytokinetic ring will form (arrows). Scale bars = 5 μm in A (applies to A,B), 2.5 μm in C,D.

Download figure to PowerPoint

thumbnail image

Figure 4. Kaz colocalizes to the cytokinetic ring. A,B: Thirty minutes after activation, kaz (arrows) clearly localizes to the cytokinetic ring. The confocal section is focused on one side of the ring. Auxiliary microtubules stemming from the main spindle also appear to come in close contact with a border of kaz (arrowheads in B). C,D: Detergent extraction revealed kaz associated with the cytoskeletal components of the ring (arrows). Asterisks are placed next to kaz-positive cytoskeletal sheets, which are major components of the detergent-extracted mammalian egg. Scale bar = 2.5 μm in A (applies to A–D).

Download figure to PowerPoint

thumbnail image

Figure 5. As development progresses, kaz reveals its dual localization. A: At 45 min after activation, a confocal image focused on the cytokinetic ring shows kaz localized to virtually the entire ring (arrows). B,C: At 60–90 min after activation, kaz staining is evident in the pronucleus, which is at the initial stages of formation (arrow). D,E: Two hours after activation kaz is clearly detected in the pronucleus (arrow). The 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining identifies the pronucleus. Scale bars = 10 μm in A (applies to A–C), 5 μm in D,E.

Download figure to PowerPoint

Most forms of kaz have an NLS (Groot et al., 2004), and antibodies to kaz have clearly revealed kaz staining within the nuclear compartment of epidermal keratinocytes. We show here that kaz staining is also prevalent within the female pronucleus as early as 1–1.25 hr after activation (Fig. 5B,C). This staining persists in well formed pronuclei (Fig. 5D,E) as well as in nuclei of blastomeres later in development (see below).

Kazrin Expression and Localization During Preimplantation Development

RT-PCR and Western blot analysis revealed kaz expression within eggs (Fig. 2), whereas confocal microscopy revealed novel localization dynamics intimately connected to egg activation (Figs. 3–5). Consequently, we analyzed kaz dynamics during early embryogenesis to further ascertain putative function(s) for kaz. Confocal analysis of kaz at the four- to eight-cell stage of embryogenesis revealed cytoplasmic/cytoskeletal and, in some blastomeres, a nuclear localization (Fig. 6). Cytoplasmic staining of kaz in detergent extracted embryos revealed partial colocalization with actin especially at cortical regions of blastomeres. Cytoskeletal sheets also appeared to be decorated with anti-kaz antibodies.

thumbnail image

Figure 6. Kaz staining in the four cell embryo partially colocalized with actin and is nuclear. A,B: Antibodies to kaz localize to the nucleus as revealed by 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining (arrowheads). Note, not all nuclei stain positively for kaz. C: Colocalized (yellow) signal was evident in blastomeres double labeled with anti-actin and anti-kaz antibodies, especially in blastomere cortical regions (arrows). D–F: Detergent extraction of four-cell embryos shows that kaz is a cytoskeletal and nuclear matrix component (arrow). Again not all nuclei stain positive for kaz. Scale bar = 10 μm in A (applies to A–F).

Download figure to PowerPoint

Kaz nuclear staining was of particular interest because it was observed only in nuclei of select blastomeres. Kaz has been shown to be present in nuclei of somatic cells; however, any function for nuclear kaz has been purely speculative. Blastomeres are in different stages of the cell cycle, and because the intensity of kaz staining varied or was absent in some blastomere nuclei, we suspected that kaz/nuclear localization may be cell cycle-dependent. As a result, we used a cell cycle analysis protocol from Hendzel et al. (1997) whereby embryos were subjected to antibodies against phosphohistone H1 and phosphohistone H3 to identify the cell cycle stage of each blastomere. Co-immunofluorescence of phosphohistone antibodies and kaz demonstrated that kaz was nuclear in blastomeres in G1, S, and G2 interphases but virtually absent during M phase beginning at prophase (before nuclear envelope breakdown) and proceeding through to telophase (Fig. 7). Consequently, kaz nuclear localization appears to be cell-cycle dependent. Of interest, we did not see kaz staining on mitotic spindles, suggesting that the kaz/meiotic spindle interaction may be unique.

thumbnail image

Figure 7. Kaz nuclear localization is cell-cycle dependent. It also becomes localized to cell–cell junction at in the blastocyst. A–C: Nuclei within the compaction stage embryo clearly reveal the loss of kaz staining near the condensed chromosomes. The technical pattern for determining cell cycle was followed by Hendzel et al. (1997). Arrows in C point to nuclei at specified cell cycle stages. D,E: At the blastocyst stage, the DNA within a trophectoderm cell (large arrow) is in prophase. The 4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) staining in E revealed that not all DNA was histone H3-positive (small arrows), suggesting that the nuclear envelope had not completely broken down; however, little or no kaz was detected in the prophase nucleus. Also present was the first detectable signal for kaz at cell–cell junctions (D; arrowheads). Only trophectoderm demonstrated this cell–cell junction phenomenon. G–I: High magnification of blastomere nuclei in the inner cell mass. Negative junctional staining, but positive nuclear staining.

Download figure to PowerPoint

It was at the blastocyst stage, however, where junctional staining for kaz was first observed. While kaz was detected in both the inner cell mass and the trophectoderm, junctional staining was only observed in the primary epithelia comprising the trophectoderm cells (Fig. 7).

Kazrin Localization After Implantation Suggests a Shift in Function

The apparent multitasking nature of kaz during fertilization and early development suggested that it may function well after preimplantation development to regulate critical postimplantation stages, including gastrulation and initial organ development. To first confirm that kaz was expressed at these stages of development, we used RT-PCR, which revealed kaz isotypes A and B within the embryo as soon as E6.0 and at E7.0 (Fig. 8). Only kaz A was detected in the ectoplacental cone (EPC) tissues at both time points (Fig. 8).

thumbnail image

Figure 8. Reverse transcriptase-polymerase chain reaction detected kaz A and B in the embryo (combined embryonic ectoderm [EE] and visceral endoderm [VE]), while detecting only kaz A in the ectoplacental cone (EPC). This pattern was similar at both embryonic day 6.0 and 7.0.

Download figure to PowerPoint

Our RT-PCR data were confirmed by confocal microscopy, which clearly revealed kaz protein expression in the E4.5 and E5.5 embryo (Figs. 9, 10). An optical section focusing on the surface of the embryo, illuminating the primitive visceral endoderm (pVE) revealed kaz staining prevalent in cell–cell junctions and within cell nuclei (Fig. 9A,B,D,F,H,I). When junctional staining for kaz was coanalyzed with DP, a measurably high colocalization was detected. DP, which has been shown to co-immunoprecipitate with kaz (Groot et al., 2004) was found primarily in extraembryonic tissues (Figs. 9, 11; Gallicanco et al., 1998, 2001). Kaz/DP localization was more readily detected and, when quantified by pixel colabeling, measured higher than kaz colocalization with another adhesion protein, E-cadherin (E-cad; Fig. 9E,F,I). That E-cad clearly colocalized with kaz introduced an intriguing idea that kaz may interact with both desmosomes and adherens junctions. The validity of the colocalization between kaz and E-cad as well as kaz and DP was supported by confocal views of DP/E-cad costaining, which showed at low and high magnification few areas of colocalization (Fig. 9G).

thumbnail image

Figure 9. Kaz localization is diverse among differentiated embryonic tissues. A–F: A whole-mount of an embryonic day (E) 5.5 day embryo was subjected to antibodies against kaz, desmoplakin (DP), and E-cadherin and visualized along the surface of the embryo by confocal microscopy. Kaz staining at cell junctions in the primitive visceral endoderm (pVE) and ectoplacental cone (EPC) tissue was clearly evident (arrowheads); however, nuclear staining was more prominent in the pVE than ectoplacental cone (EPC). C,D: Costaining using anti-kaz and anti-DP revealed marked overlap of DP and kaz (yellow signal). E,F: Costaining using anti-kaz and anti–E-cadherin also revealed some colocalization, but at decreased levels when compared with DP. G: Low- and high-magnification view (shown) of embryos colabeled with anti-DP (arrows) and anti E-cad (arrowhead) revealed little colocalization and provided evidence that the DP/kaz and E-cad/kaz colocalization signals are specific. H,I: Quantification of pixel overlap confirm confocal observations. H: Far more kaz is in pVE nuclei than EPC nuclei. The control DP/4′,6-diamidine-2-phenylidole-dihydrochloride (DAPI) colocalization showing virtually no signal confirms that the kaz/DAPI signal is authentic. No significant difference is observed between EPC and pVE with regard to kaz/DP colocalization. Although the kaz/E-cad colocalization is ∼2 times lower, the fact that a signal is detectable suggests that kaz may be intimately associated with adherens junctions. Scale bar = 25 μm in A (applies to A–F).

Download figure to PowerPoint

thumbnail image

Figure 10. Confocal microscopy revealed kaz primarily in extraembryonic tissues. A–D: An optical section through the middle of a whole-mount embryonic day 5.5 day embryo shows kaz residing primarily in the primitive visceral endoderm (pVE) and extraembryonic ectoderm (EXE). Faint kaz staining is present in the primitive embryonic ectoderm (pEE), but it was mostly cytoplasmic and nuclear. E-cadherin staining in C serves as a positive control for all cell types at this stage, whereas desmoplakin (DP) staining (arrowheads) isolates the extraembryonic tissues (Gallicano et al., 1998, 2001). Asterisks, proamniotic cavity. Scale bar = 25 μm in A (applies to A–D).

Download figure to PowerPoint

thumbnail image

Figure 11. Kaz localization is altered in desmoplakin (DP)−/− embryos A: Using antibodies against E-cadherin, an optical section through an embryonic day 4.5 embryo reveals junctional staining between both inner cell mass (ICM) and trophectoderm (Troph; arrows). B: In the absence of DP, kaz staining in the cytoplasm (arrows) appears markedly higher than in wild-type embryos containing DP (compare with Figs. 9 and 10). Some junctional staining of kaz persists (arrowheads). High-magnification views of B (B1 and B2) clearly show junctional staining (arrowheads in B1) and the random dispersal pattern (arrow in B2) of kaz throughout the cell. C: No DP staining is detected in this embryo. D: Double labeling of kaz and E-cadherin strengthen the premise that kaz can associate with adherens junctions (arrowheads); however, areas of non colabeling at cell–cell junctions suggest that kaz may be able to bind other junctional components.

Download figure to PowerPoint

Kaz staining in the E5.5 EPC tissue was predominantly at cell–cell junctions. Quantification of kaz colocalization with DP at cell–cell junctions in the EPC was not significantly different than at junctions in cells residing in the pVE. However, nuclear staining for kaz in EPC tissue was dramatically reduced when compared with nuclear localization in pVE cells (Fig. 9H). Nuclear colocalization in pVE cells, however, was 4.7 times higher than cells populating the EPC. Of 48 pVE nuclei from three embryos counted by direct observation or by the Fluoview colocalization view-processor, 38 nuclei (79%) were positive for kaz (Fig. 9; nuclei were considered positive if the pixel intensity signal for kaz was above 25 of 255 or was visually clear). In contrast, of 52 nuclei counted in the EPC, only 24 were positive (46%) for kaz suggesting that kaz nuclear function is different among embryonic and extraembryonic tissue types.

An optical section through the middle of the E5.5 embryo revealed prominent kaz staining primarily within extraembryonic tissues including the pVE, and the extraembryonic ectoderm (EXE). Kaz staining was observed in the primitive embryonic ectoderm (pEE); however, the level of staining was markedly lower than extraembryonic tissues (Fig. 10). The paucity of kaz staining in the pEE was primarily cytoplasmic and nuclear. Very little junctional or nuclear staining was detectable. One reason for the lack of junctional staining maybe because desmosomes (and DP) are not present in pEE cells at this stage of development.

To further understand the relevance of the kaz/DP colocalization, we examined DP−/− embryos using confocal microscopy. E 4.5 DP−/− embryos (Fig. 11; as attested to by a complete loss of DP staining) revealed an observed increase in cytoplasmic staining of kaz when compared with DP+ embryos. Without DP, the close localization of kaz with E-cad also was more evident; however, the number of kaz/E-cad colabeling areas in DP−/− embryos did not appear to increase over kaz/E-cad staining in DP+ embryos.

By E7, gastrulation is well under way. The importance of this stage is the transition of the EE through the primitive streak to form the three germ layers: ectoderm, endoderm, and mesoderm (Tam et al., 1993). Confocal microscopy clearly revealed a highly complex staining pattern for kaz as development progressed (Fig. 12). In the growing epithelial region of the E7 EPC, strong kaz staining was consistently observed at cell junctions in the cytotrophoblast cells. These cells also demonstrated prevalent DP staining at cell–cell junctions, much of which colocalized with kaz (Fig. 12A,C,D,F). Colocalization between kaz and DP was virtually absent within syncytiotrophoblasts (Fig. 12B,C).

thumbnail image

Figure 12. A–F: A clear divergence in staining is observed between the extraembryonic mesoderm and the tissue types populating the ectoplacental cone (EPC). A,C,D,F: Low- and corresponding high-magnification views of the cytotrophoblast (cyt) region of the EPC reveals prominent desmoplakin (DP) staining at cell–cell junctions (arrows). Strong DP staining is present in the syncytial trophoblast cells (syn); however, virtually no DP is present in the extraembryonic mesoderm (EXM). B,C,E,F: In cytotrophoblast cells, some kaz nuclear staining is evident; however, kaz is primarily junctional colocalizing with DP. In syn cells, kaz staining is virtually absent. In EXM cells, kaz is primarily nuclear, whereas very little staining is detected at cell–cell contacts. Scale bars = 50 μm in A (applies to A–C), 10 μm in D,E.

Download figure to PowerPoint

Cells populating the extraembryonic mesoderm (EXM) lining the exocoelomic cavity also were positive for kaz; however, upon close inspection of these cells (Fig. 12E,F), the majority of the kaz signal in the EXM was nuclear. Of interest, DP and desmosome staining in these cells were negative, which may explain the increased nuclear staining; that is, without DP to link to, kaz may interact with nuclear components with higher affinity.

More intriguing was the localization of kaz within anterior embryonic ectoderm in the region comprising presumptive neurectoderm. Kaz was found in distinct subpopulations of cells (Fig. 13). Colocalization with DP was evident; however, the most intense nuclear signal for kaz at any stage in early development was clearly recorded within these islands of cells (Fig. 13B,C). Imaging these cells by differential interference contrast microscopy showed no apparent morphological differences between the distinct kaz-positive cells and surrounding kaz-negative cells (Fig. 13D). These results demonstrate for the first time the existence of islands of EE cells that contain a protein, kaz, capable of interacting with nuclear components (e.g., nuclear matrix, etc.) and desmosomes and adherens junctions.

thumbnail image

Figure 13. An optical section traversing the embryonic ectoderm (EE) of an embryonic day (E) 7.5 embryo reveals islands of cells positive for kaz. A–C: Most cells are positive for desmoplakin (DP; A); however, kaz (B) is primarily found in islands of cells. Its localization is nuclear (arrowheads) and junctional (arrows) in EE cells with junctional staining colocalizing primarily with DP (C) and not E-cadherin (data not shown). D: Morphologically, kaz-positive cells (arrows and arrowheads) look no different than neighboring kaz-negative cells. Scale bar = 10 μm in A (applies to A–D).

Download figure to PowerPoint

Kaz staining at the primitive streak was restricted to nuclear staining of mesoderm cells exiting and moving away from the streak (data not shown). Little or no staining was observed in the EE cells populating the remaining primitive streak.

As shown in Figure 12, the EXM component bordering the exocoelomic cavity contained kaz but not desmosomes. Moreover, we speculated that the increased nuclear signal was due to the lack of DP for kaz to link to. To assess this possibility, we used confocal microscopy to visualize kaz and DP within the developing heart of E9.0 embryos (Fig. 14). The heart is a mesodermally derived organ that contains desmosomes beginning at about E8.5 (van der Loop et al., 1995; Arai et al., 1997). While concentrated areas with directly overlapping kaz and DP signals were present, we unexpectedly found that the majority of signal was either nuclear, cytoplasmic, or at cell–cell junctions not associated with DP. Consequently, nuclear localization within mesoderm cells and tissue may be inherent to that tissue type and not due to the presence of DP.

thumbnail image

Figure 14. Kaz and desmoplakin (DP) do not overtly colocalize in the developing heart. A: A confocal view of the ventricle of an embryonic day (E) 9.0 day heart shows significant DP staining (arrows) within presumptive desmosomes. B: Kaz localizes to nuclei and cytoplasmic domains within cardiomyocytes. C: Colocalization of DP and kaz antibodies reveals little colocalization. Arrows point to areas of nonoverlap. Epi, epicardium; myo, myocardium; endo, endocardium; ant, anterior axis. Scale bar = 100 μm in A (applies to A–C).

Download figure to PowerPoint

DISCUSSION

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

We show, for the first time, the existence of a plakin binding protein, kaz, within mammalian eggs and preimplantation and postimplantation embryos. In eggs, kaz localized specifically to the meiotic spindle; however, it showed clear dynamic changes in localization as soon as 10 min after egg activation. Confocal microscopy demonstrated that kaz is likely involved in cytokinetic events of second polar body emission most likely interacting with the cytoskeletal components of the ring (i.e., actin, actin-associated proteins, etc.). This transient staining pattern was followed by distinct pronuclear/nuclear matrix localization as soon as 2 hr after egg activation. Also evident was its localization to cytoskeletal sheets, which are major cytoskeletal components in mammalian eggs and early embryos. As development progressed, kaz colocalized with the actin filament network in blastomeres and in postcompaction embryos especially in cortical regions of blastomeres. Nuclear staining of kaz persisted through early development; however, this pattern of staining was clearly linked to the cell cycle. The signal was virtually absent when blastomeres entered and proceeded through M-phase, suggesting that, during this phase of the cell cycle, kaz does not directly associate with nuclear matrix components or components associated with condensed chromosomes.

Kaz dynamics in postimplantation embryos appear to be extremely complex. At E5.5, the majority of kaz staining was found within extraembryonic tissues, with some minimal staining in the primitive ectoderm (the tissue that will form the mouse proper). Whereas, in general, the staining in extraembryonic tissues appeared both junctional and nuclear, divergences in kaz localization were detected among various extraembryonic tissue types. For example, the number of nuclei and the average intensity of kaz within the nuclear compartment of cells comprising the pEE was markedly higher than cells comprising the EPC.

Analysis of junctional kaz revealed that it preferentially colocalized with DP over another junctional protein, E-cad; however, some colocalization with E-cadherin was evident at these early stages (E5.5–E7.5) of implantation. In fact, DP−/− embryos clearly showed kaz closely associated, and in some cases colocalized, with E-cadherin. Whereas it was clearly demonstrated in keratinocytes that kaz colocalized and co-immunoprecipitated with DP, ample evidence also revealed that kaz colocalized with cortical actin especially near microvilli (Groot et al., 2004). Our data appear to confirm and advance the suggestion of Groot et al. (2004) that kaz may be involved in the interplay between adherens junctions and desmosomes, especially in embryonic tissues that differentiate into primitive epithelia and form the cytotrophoblast portion of the placenta and yolk sac. However, a function for the E-cadherin association remains to be determined.

Although our data tend to support the suggestion that kaz may link plakins within desmosomes with actin filaments near adherens junctions, at least later in development, during preimplantation development, desmosomes and adherens junctions are not present. Adherens junctions arise during embryogenesis at the 8- to 16-cell stage during compaction. Desmosome formation follows shortly thereafter (Ducibella et al., 1975; Jackson et al., 1980; Fleming et al., 1991; Gallicano et al., 1998). It is tempting to propose that kaz/actin binding at preimplantation stages may be a prerequisite for guiding actin to newly forming adherens junction. Experiments are proposed to test this hypothesis.

The significance of kaz expression in eggs and embryos resides on the evidence that kaz undergoes dynamic, spatial reconfiguration at critical developmental transitions within two domains, one that is cytoskeletal and one that is nuclear. Developmental transitions are predictable events that must happen sequentially during embryogenesis in order for the embryo to successfully complete the developmental process. Some transitions are common to most phyla (e.g., fertilization), whereas others are mammalian specific (e.g., compaction, implantation, germinal disk/egg cylinder formation). At the molecular level, the most common feature to each transition is a complete remodeling of specific cytoskeletal components. In fact, decades of research have explicitly shown that the cytoskeleton is a driving force for each transition (Ducibella et al., 1975; Schatten et al., 1985; Bement et al., 1992; Riethmacher et al., 1995; Clayton et al., 1999; Pauken and Capco, 1999; Gallicano, 2001, and references therein). Furthermore, each cytoskeletal system (i.e., actin, microtubules, and intermediate filaments) has adapted specialized features to allow them to push through the developmental transition. For example, at fertilization, a loose actin-based network of filaments intimately associates with the head and neck of a sperm for normal, efficient sperm penetration; a specialized function for actin that occurs only once in the lifetime of an organism (Maro et al., 1985; Le Guen at al., 1989; Webster and McGaughey, 1990). Our data suggest that kaz also has adapted a specialized role during these early transitions.

When proteins such as kaz are observed altering their location concomitantly with certain developmental transitions, one conclusion is that kaz is involved in facilitating the successful completion of the transition. CamKII is a prime example of this premise. Its localization specifically changes spatially and temporally to drive second polar body emission (Winston and Maro, 1995; Johnson et al., 1998). Consequently, based on the kaz data gathered thus far, we propose that kaz is likely a key component that associates with the cytoskeleton to coordinate the mechanism(s) that drive specific developmental transitions, including second polar body, compaction, and possibly postimplantation development.

However, the cytoskeletal characteristic of kaz during development is only half the story. Its association within the nucleus, more specifically the detergent-resistant nuclear matrix, makes it tempting to suggest that kaz acts within some aspect of gene regulation, nuclear matrix stability, or both. We found kaz/nuclear staining to be associated with the cell cycle. Its virtual disappearance prior to M phase and its appearance at the initial stages of interphase suggested that it is a structural component of nuclear matrix assembly/disassembly regulatory mechanism (Martens et al., 2002; Stein et al., 2004). Proteins such as p300, a transcriptional coactivator that associates with SMADs and STATs, act by binding to specific nuclear matrix components, which, in turn, bind promoter elements within genes at specific times to direct proliferation and/or differentiation (Martens et al., 2002). Given the ability of kaz to link cytoskeletal systems, its leucine zipper-like motif, and its NLS, we propose that kaz may be a critical protein involved in nuclear matrix stability, regulation, and/or gene regulation.

Suggesting that kaz may have a dual function, one nuclear, the other cytoskeletal/adhesive, is not a foreign concept. Dinsmore and Sloboda (1988) and Johnston and Sloboda (1992) revealed a 62-kD protein that associated with the mitotic apparatus during M phase and the nucleus during interphase. Basonuclin, a zinc-finger transcription factor found associated with rRNA machinery in nuclei of primary spermatocytes was also found to localize to the basal body, a specialized centriolar component of the developing sperm tail (Yang et al., 1997; Mahoney et al., 1998). Of course, the armadillo (arm) family, which includes β-catenin and plakoglobin has been shown to undoubtedly carry out specific actions both in the nucleus and at cell junctions (Aberle et al., 1996; Hecht et al., 1999; Hu et al., 2003). As a result, when the data for kaz presented here is combined with examples of dual function of other cytolinker proteins, the evidence becomes clear that kaz most likely has a multitasking function.

We suspect that kaz is active when in the nucleus. This supposition stems from findings within E5.5 embryos where almost twice as many nuclei of pEE cells contain kaz compared with nuclei of EPC cells. This observation may be critical because other arm, dual-action proteins share this nonrandom nuclear localization when the protein has a definite nuclear function. The obvious example is the staining pattern for β-catenin where, in postimplantation embryos, it clearly localizes to more nuclei in EE cells at the primitive streak than nuclei of surrounding nonprimitive streak EE cells (Mohamed et al., 2004). Conversely, arm proteins such as plakophilins, show a broader nuclear staining pattern and, correspondingly, do not appear to have a nuclear function early in development (i.e., fertilization–E7.5; Schmidt et al., 1999; South et al., 2003; Grossmann et al., 2004).

More conclusive evidence that kaz may function within the nucleus is seen in E7.5 embryos where distinct islands of kaz-positive EE cells contained markedly high levels of nuclear kaz. The location of these cells places them within the developing neural plate. It is tempting to suggest that these specialized cells may contribute to the development of the presumptive neurectoderm. Future investigation using siRNA to knockdown kaz in eggs and early embryos will be necessary to determine its role(s) especially within the nuclear matrix.

EXPERIMENTAL PROCEDURES

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

Egg and Embryo Procurement

Female FVB mice were gonadotropin primed. Eggs arrested at meiotic metaphase II were collected from oviducts with a modified version of previously described protocols (Capco and McGaughey, 1986; McGaughey and Capco, 1989; Gallicano and Capco, 1995). Eggs were transferred to organ culture dishes (No. 3037, Falcon, Inc., Lincoln Park, NJ) no later than 15 hr after injection with human chorionic gonadotropin (hCG) in prewarmed FHM medium (Specialty Media, Phillipsburg, NJ). Cumulus cells surrounding eggs were removed by incubation in 0.1% (w/v) hyaluronidase (Sigma, Inc., St. Louis, MO) in culture medium.

Egg activation was initiated by incubation in 7% ethanol (EtOH) in FHM medium supplemented with 0.3% bovine serum albumin and cultured for 5 min at 35°C before being washed in EtOH-free medium (Markoulaki et al., 2004). Eggs were examined at 10, 20, 30, 45, 60, 90, and 120 min.

To obtain embryos, hormone injections were timed similarly as those for acquiring eggs, except 3 hr after hCG injection, one female was placed into a cage with one male to allow mating. One-, four-, and eight-cell embryos were obtained 10–12, 48, and 60 hr after fertilization, respectively, by teasing oviducts with Dumont No. 5 forceps. Isolation of postimplantation embryos was performed by following procedures found in Nagy et al. (2003).

RT-PCR

Eggs, embryos, or mouse embryonic fibroblasts (MEFs; as controls) used for RT-PCR were briefly washed in DEPC phosphate buffered saline (PBS), and immediately placed into Trizol reagent (Life Technologies, Inc., Gaithersburg, MD) to isolate RNA. Yeast tRNA was added as a carrier. RT-PCR was carried out using the Onestep RT-PCR kit (Life Technologies, Inc.). Two microliters of RNA were pipetted into a 20-μl RT-PCR reaction mixture containing 1× RT-PCR buffer, 40 ng of each primer, optimal concentration of MgCl2, 0.2 mM dNTPs, and 2 U of reverse transcriptase and Taq polymerase. The DNA template was denatured at 94°C, followed by amplification using the following parameters: 94°C for 20 sec, 56°C for 20 sec, and 72°C for 30 sec, for 40 cycles. Control RT-PCR reactions were run for each batch of mRNA. First, no reverse transcriptase was used in concurrent reactions to determine if contaminating DNA was present. Second, tubulin primers were used as a positive control for all PCR reactions. Primer sequences for Kaz were obtained from Groot et al. (2004). Ten-microliter aliquots of each PCR product were electrophoresed on a 2% agarose gel and visualized using ethidium bromide.

Western Blot Analysis

A total of 125 eggs, embryos, or equal quantity of MEF protein were immediately placed into separate Eppendorf tubes containing sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS w/v, 1 mM β-mercaptoethanol, 10% glycerol). Samples were boiled for 5 min and loaded onto 10% polyacrylamide gels with molecular weight markers and separated by SDS-PAGE. Proteins were transferred to PVDF membrane and blocked with 2.0% bovine serum albumin in PBS, pH 7.4, with 0.1% Tween 20 at room temperature (rt). Blots were challenged with primary antibody at dilutions of 1/500–1/1,000 in block overnight at 4°C, followed by washing 3× with PBT (1× PBS + 0.1% Tween 20) at rt and challenged with appropriate secondary antibody conjugated to horseradish peroxidase (Pierce, Inc. Rockford, IL). As a positive control, tissue samples known to contain kaz were always loaded in a subsequent lane to verify the efficacy of the anti-kaz used for each experiment.

For analysis of the detergent-resistant cytoskeleton, we modified a procedure from Gallicano et al. (1991). Briefly, detergent-soluble components were collected and precipitated in four volumes of ethanol precooled to −20°C. After precipitation, the ethanol was removed, the pellet dried, and the precipitate dissolved in Laemmli sample buffer (Laemmli, 1970). The remaining detergent-insoluble pellet was directly dissolved in Laemmli sample buffer. Electrophoresis was performed as described above. Antibodies used were anti-kaz (a kind gift of Dr. Fiona Watt) and anti-tubulin (Sigma).

Confocal Microscopy and Analysis

Confocal analysis was conducted on eggs and embryos previously fixed in 4% paraformaldehyde for 1 hr, followed by permeabilization with 1% Triton-X-100 for 30 min. After two washes in PBS, specimens were blocked with a protein mixture from the M.O.M. Kit from Vector Laboratories, Inc. (Burlingame, CA). Primary and secondary antibodies were diluted in blocking solution, and eggs/embryos were incubated with each antibody for 1 hr at 37°C. In the final step, eggs/embryos were moved directly from secondary antibody into a DAPI solution at 2 μg/ml and mounted onto slides with anti-fade and sealed with coverslips. Mounting eggs and embryos onto slides was accomplished using the technique described by Gallicano and Capco (1995). Detergent extraction of eggs and embryos before fixation was conducted using the protocol from Gallicano et al. (1991, 1992).

Confocal images were acquired using an Olympus Fluoview 500 Laser Scanning Microscope (Olympus America, Inc., Melville, NY) using the accompanying Fluoview image acquisition and analysis software (version 4.3). Cells within areas were imaged using a 1.4 numerical aperture, ×60 Olympus objective. Antibodies used were as follows: anti-kaz, anti-tubulin, anti-desmoplakin (Zhou et al., 2004), anti-actin (Sigma), and anti-E-cadherin (Santa Cruz, Inc., CA).

To quantify colocalization of kaz with other components, images were collected using the Olympus Fluoview 500 Confocal Laser Scanning Microscope (Olympus America, Inc., Melville, NY). Acquired images were analyzed by using the Olympus Fluoview colocalization view-processor to produce annotations showing overlapping pixels. Original images and annotations were then imported into MetaMorph (Universal Imaging Co., Downingtown, PA) to quantify the colocalized pixels. Regions of interest were first chosen on the original data and then transferred to the annotations showing colocalized pixels. Measurement of region statistics in MetaMorph gave a count of the colocalized pixels contained in each selected region.

Cell cycle analysis was performed using a protocol similar to Hendzel et al., (1997). The protocol can be found at http://www.upstate.com/misc/protocols/asp?prot=cellcyclestag from Upstate, Inc.

Acknowledgements

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

We thank Dr. Fiona Watt for the kazrin antiserum and Dr. Lisa Sevilla for her technical help with our questions about kazrin. We also thank Dr. Bob Lechleider and Tammy Gallicano for their critical reading of the manuscript. This work was supported by a grant from NHLBI at the NIH, grant #HL70204.

REFERENCES

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