Correlative microscopy and block-face imaging (CoMBI): a 3D imaging method with wide applicability in the field of biological science.

Correlative microscopy and block-face imaging (CoMBI) is an imaging method, which is characterized by the ability to obtain both serial block-face images as a 3-dimentional (3D) dataset and sections for 2-dimentional (2D) light microscopic analysis. These 3D and 2D morphological data can be correlated with each other to facilitate data interpretation. CoMBI is an easy-to-install and low-cost 3D imaging method since its system can be assembled by the researcher using a regular microtome, consumer digital camera, and some self-made devices, and its installation and instruction manuals are open-source. After the first release of CoMBI method from our laboratory, CoMBI systems have been installed in more than a dozen laboratories and are used for 3D analysis of various biological specimens. Typical application of CoMBI is 3D anatomical analysis using the natural color and contrast of the specimen. We have been using CoMBI for analyzing human brain to obtain the fine 3D anatomy as a reference to determine the causes of neurological diseases and to improve the effectiveness of surgery. Recently, we have been using CoMBI for detecting the colors of chromogens, which are used for labeling specific molecules. Mouse embryos colored with X-gal, a conventional chromogen for detecting LacZ products, were imaged using CoMBI, and the 3D distribution of X-gal was successfully visualized. Thus, CoMBI can now be used for many purposes, including 3D anatomical analysis, 2D microscopy using sections, and 3D distribution of specific molecules. These suggest that CoMBI should be more widely used in the field of biological research.

Recent technological advances in correlative light and electron microscopy for the comprehensive analysis of neural circuits.

Light microscopy (LM) covers a relatively wide area and is suitable for observing the entire neuronal network. However, resolution of LM is insufficient to identify synapses and determine whether neighboring neurons are connected via synapses. In contrast, the resolution of electron microscopy (EM) is sufficiently high to detect synapses and is useful for identifying neuronal connectivity; however, serial images cannot easily show the entire morphology of neurons, as EM covers a relatively narrow region. Thus, covering a large area requires a large dataset. Furthermore, the three-dimensional (3D) reconstruction of neurons by EM requires considerable time and effort, and the segmentation of neurons is laborious. Correlative light and electron microscopy (CLEM) is an approach for correlating images obtained via LM and EM. Because LM and EM are complementary in terms of compensating for their shortcomings, CLEM is a powerful technique for the comprehensive analysis of neural circuits. This review provides an overview of recent advances in CLEM tools and methods, particularly the fluorescent probes available for CLEM and near-infrared branding technique to match LM and EM images. We also discuss the challenges and limitations associated with contemporary CLEM technologies.

Correlative microscopy and block-face imaging (CoMBI) method for both paraffin-embedded and frozen specimens.

Correlative microscopy and block-face imaging (CoMBI), a method that we previously developed, is characterized by the ability to correlate between serial block-face images as 3-dimensional (3D) datasets and sections as 2-dimensional (2D) microscopic images. CoMBI has been performed for the morphological analyses of various biological specimens, and its use is expanding. However, the conventional CoMBI system utilizes a cryostat, which limits its compatibility to only frozen blocks and the resolution of the block-face image. We developed a new CoMBI system that can be applied to not only frozen blocks but also paraffin blocks, and it has an improved magnification for block-face imaging. The new system, called CoMBI-S, comprises sliding-type sectioning devices and imaging devices, and it conducts block slicing and block-face imaging automatically. Sections can also be collected and processed for microscopy as required. We also developed sample preparation methods for improving the qualities of the block-face images and 3D rendered volumes. We successfully obtained correlative 3D datasets and 2D microscopic images of zebrafish, mice, and fruit flies, which were paraffin-embedded or frozen. In addition, the 3D datasets at the highest magnification could depict a single neuron and bile canaliculus.

Loss of VAMP5 in mice results in duplication of the ureter and insufficient expansion of the lung.

BACKGROUND: Vesicle-associated membrane protein 5 (VAMP5) is a member of the SNARE protein family, which regulates the docking and fusion of membrane vesicles within cells. Previously, we reported ubiquitous expression of VAMP5 proteins in various organs except the brain and small intestine. However, the precise roles of VAMP5 in each organ remain unclear. To explore the roles of VAMP5 in vivo, we generated VAMP5 knockout (KO) mice. RESULTS: VAMP5 KO mice showed low birth rate and low body weight. KO embryos grew normally in the uterus, and tended to die around birth. Anatomical analysis revealed that viable KO mice often exhibited duplication of the ureter, and dead KO mice showed insufficient expansion of the lung. VAMP5 was localized in the epithelial cells of the ureter and terminal bronchiole. CONCLUSIONS: VAMP5 KO mice showed a low birth rate and abnormalities of the urinary and respiratory systems. VAMP5 KO mice died around birth, possibly due to defects in vesicoureteral flow and breathing. The results presented could provide a basis for future studies to understand the roles of VAMP5 protein. Developmental Dynamics 247:754-762, 2018. © 2018 Wiley Periodicals, Inc.

A novel imaging method for correlating 2D light microscopic data and 3D volume data based on block-face imaging.

We have developed an imaging method designated as correlative light microscopy and block-face imaging (CoMBI), which contributes to improve the reliability of morphological analyses. This method can collect both the frozen sections and serial block-face images in a single specimen. The frozen section can be used for conventional light microscopic analysis to obtain 2-dimensional (2D) anatomical and molecular information, while serial block-face images can be used as 3-dimensional (3D) volume data for anatomical analysis. Thus, the sections maintain positional information in the specimen, and allows the correlation of 2D microscopic data and 3D volume data in a single specimen. The subjects can vary in size and type, and can cover most specimens encountered in biology. In addition, the required system for our method is characterized by cost-effectiveness. Here, we demonstrated the utility of CoMBI using specimens ranging in size from several millimeters to several centimeters, i.e., mouse embryos, human brainstem samples, and stag beetle larvae, and present successful correlation between the 2D light microscopic images and 3D volume data in a single specimen.

The correlation between femoral sulcus morphology and osteoarthritic changes in the patello-femoral joint.

PURPOSE: The purpose of this study was to reveal the correlation between the type of lesion and the depth of osteoarthritic (OA) changes in the patello-femoral (PF) joint and its bony morphological characteristics using computed tomography (CT) data. METHODS: Eighty-seven cadaveric knees were included in this study with median age of 83 years (62-97). OA depth evaluation was performed following Outerbridge's classification. Patella OA lesions were classified macroscopically using Han's method: type (1) no or minimal lesion, type (2) medial facet lesion without involvement of the ridge, type (3) lateral facet lesion without involvement of the ridge, type (4) lesion involving the ridge only, type (5) medial facet lesion with involvement of the ridge, type (6) lateral facet lesion with involvement of the ridge, and type (7) global lesion. Femoral-side OA lesions in the PF joint were classified using a modified Chang's method. Type (1) no or minimal lesion, type (2) medial facet lesion, type (3) centre of patella groove lesion, type (4) lateral facet lesion, and type (5) global lesion. Whole-body CTs of all cadavers were taken before knee dissection. Using the CT data, patella morphology was evaluated following Wiberg's classification. Femoral sulcus angle (SA), sulcus depth (SD), and sulcus width (SW) were also measured using CT data. RESULTS: The measured SA, SD, and SW were 144.8° ± 7.2°, 7.0 ± 1.6 mm and 3.4 ± 0.3 mm, respectively. When patella OA depth was divided into grades 1-2 (n = 30) and grades 3-4 (n = 57), the SD of grade 1-2 knees was 6.5 ± 1.3 mm, and the SD of grade 3-4 knees was 7.3 ± 1.6 mm, constituting a significant difference (p = 0.01). No significant difference in either SA or SW was observed between the two groups. Patella OA lesion, femoral-side OA lesion, and depth were not affected by SA, SD, or SW. Wiberg's classification also showed no significant correlation with PF-OA. CONCLUSION: Deep SD was significantly correlated with the incidence of severe patella OA. Wiberg's classification, SA, and SW were not correlated with PF-OA. For clinical relevance, there is a risk of PF-OA progression in patients with deep SD, and treatment should be applied accordingly.

3D lung motion assessments on inspiratory/expiratory thin-section CT: Capability for pulmonary functional loss of smoking-related COPD in comparison with lung destruction and air trapping.

PURPOSE: To evaluate the utility of three-dimensional (3D) lung motion on inspiratory and expiratory CT for pulmonary functional loss in smoking-related COPD in comparison with lung destruction and air trapping assessments. METHOD AND MATERIALS: Forty-four consecutive smokers and COPD patients prospectively underwent inspiratory and expiratory CT. A 3D motion vector map was generated from these CTs, and regional motion magnitudes were measured at the horizontal axis (X-axis), the ventrodorsal axis (Y-axis), and the craniocaudal axis (Z-axis). All mean magnitudes within the entire lung (MMLX, MMLY, and MMLZ) were normalized by expiratory CT lung volume. Moreover, CT-based functional lung volume (FLV) on inspiratory CT and air trapping lung volume (ATLV) on expiratory CT were assessed quantitatively. To evaluate the capability for pulmonary function loss assessment, all MMLs were correlated with pulmonary function tests. Then, discrimination analysis was performed to determine the concordance capability for clinical stage, and correct classification capabilities were compared by means of McNemar's test. RESULTS: Multiple regression analysis showed MMLY (ß=0.657, p<0.001) and FLV (ß=0.375, p=0.019) were correlated with percentage of predicted forced expiratory volume in 1 second. Correct classification capabilities using patient characteristics and MMLs (68.2 (30/44)%) were significantly higher than those obtained by patient characteristics, FLV, and ATLV (54.5 (24/44)%), p=0.031). CONCLUSION: 3D lung motion parameter assessment is useful for smoking-related COPD assessment as well as lung parenchymal destruction and/or air trapping evaluations.

Organization of organelles and VAMP-associated vesicular transport systems in differentiating skeletal muscle cells.

Vesicular transport plays an important role in the regulation of cellular function and differentiation of the cell, and intracellular vesicles play a role in the delivery of membrane components and in sorting membrane proteins to appropriate domains in organelles and the plasma membrane. Research on vesicular transport in differentiating cells has mostly focused on neurons and epithelial cells, and few such studies have been carried out on skeletal muscle cells. Skeletal muscle cells have specialized organelles and plasma membrane domains, including T-tubules, sarcoplasmic reticulum, neuromuscular junctions, and myotendinous junctions. The differentiation of skeletal muscle cells is achieved by multiple steps, i.e., proliferation of myoblasts, formation of myotubes by cell-cell fusion, and maturation of myotubes into myofibers. Systematic vesicular transport is expected to play a role in the maintenance and development of skeletal muscle cells. Here, we review a map of the vesicular transport system during the differentiation of skeletal muscle cells. The characteristics of organelle arrangement in myotubes are described according to morphological studies. Vesicular transport in myotubes is explained by the expression profiles of soluble N-ethylmaleimide-sensitive factor attachment protein receptor proteins.

Filamentous structures in skeletal muscle: anchors for the subsarcolemmal space.

In skeletal muscle fibers, intermediate filaments and actin filaments provide structural support to the myofibrils and the sarcolemma. For many years, it was poorly understood from ultrastructural observations that how these filamentous structures were kept anchored. The present study was conducted to determine the architecture of filamentous anchoring structures in the subsarcolemmal space and the intermyofibrils. The diaphragms (Dp) of adult wild type and mdx mice (mdx is a model for Duchenne muscular dystrophy) were subjected to tension applied perpendicular to the long axis of the muscle fibers, with or without treatment with 1% Triton X-100 or 0.03% saponin. These experiments were conducted to confirm the presence and integrity of the filamentous anchoring structures. Transmission electron microscopy revealed that these structures provide firm transverse connections between the sarcolemma and peripheral myofibrils. Most of the filamentous structures appeared to be inserted into subsarcolemmal densities, forming anchoring connections between the sarcolemma and peripheral myofibrils. In some cases, actin filaments were found to run longitudinally in the subsarcolemmal space to connect to the sarcolemma or in some cases to connect to the intermyofibrils as elongated thin filaments. These filamentous anchoring structures were less common in the mdx Dp. Our data suggest that the transverse and longitudinal filamentous structures form an anchoring system in the subsarcolemmal space and the intermyofibrils.

Identification of a naturally occurring retinoid X receptor agonist from Brazilian green propolis.

BACKGROUND: Brazilian green propolis (BGP), a resinous substance produced from Baccharis dracunculifolia by Africanized honey bees (Apis mellifera), is used as a folk medicine. Our present study explores the retinoid X receptor (RXR) agonistic activity of BGP and the identification of an RXR agonist in its extract. METHODS: RXRα agonistic activity was evaluated using a luciferase reporter gene assay. Isolation of the RXRα agonist from the ethanolic extract of BGP was performed using successive silica gel and a reversed phase column chromatography. The interaction between the isolated RXRα agonist and RXRα protein was predicted by a receptor-ligand docking simulation. The nuclear receptor (NR) cofactor assay was used to estimate whether the isolated RXRα agonist bound to various NRs, including RXRs and peroxisome proliferator-activated receptors (PPARs). We further examined its effect on adipogenesis in 3T3-L1 fibroblasts. RESULTS: We identified drupanin as an RXRα agonist with an EC50 value of 4.8 ± 1.0 µM. Drupanin activated three RXR subtypes by a similar amount and activated PPARγ moderately. Additionally, drupanin induced adipogenesis and elevated aP2 mRNA levels in 3T3-L1 fibroblasts. CONCLUSIONS: Drupanin, a component of BGP, is a novel RXR agonist with slight PPARγ agonistic activity. GENERAL SIGNIFICANCE: This study revealed for the first time that BGP activates RXR and drupanin is an RXR agonist in its extract.

An integrated teaching method of gross anatomy and computed tomography radiology.

It is essential for medical students to learn and comprehend human anatomy in three dimensions (3D). With this in mind, a new system was designed in order to integrate anatomical dissections with diagnostic computed tomography (CT) radiology. Cadavers were scanned by CT scanners, and students then consulted the postmortem CT images during cadaver dissection to gain a better understanding of 3D human anatomy and diagnostic radiology. Students used handheld digital imaging and communications in medicine viewers at the bench-side (OsiriX on iPod touch or iPad), which enabled "pixel-to-tissue" direct comparisons of CT images and cadavers. Students had lectures and workshops on diagnostic radiology, and they completed study assignments where they discussed findings in the anatomy laboratory compared with CT radiology findings. This teaching method for gross and radiological anatomy was used beginning in 2009, and it yielded strongly positive student perspectives and significant improvements in radiology skills in later clinical courses.

Vesicular transport system in myotubes: ultrastructural study and signposting with vesicle-associated membrane proteins.

Myofibers have characteristic membrane compartments in their cytoplasm and sarcolemma, such as the sarcoplasmic reticulum, T-tubules, neuromuscular junction, and myotendinous junction. Little is known about the vesicular transport that is believed to mediate the development of these membrane compartments. We determined the locations of organelles in differentiating myotubes. Electron microscopic observation of a whole myotube revealed the arrangement of Golgi apparatus, rough endoplasmic reticulum, autolysosomes, mitochondria, and smooth endoplasmic reticulum from the perinuclear region toward the end of myotubes and the existence of a large number of vesicles near the ends of myotubes. Vesicles in myotubes were further characterized using immunofluorescence microscopy to analyze expression and localization of vesicle-associated membrane proteins (VAMPs). VAMPs are a family of seven proteins that regulate post-Golgi vesicular transport via the fusion of vesicles to the target membranes. Myotubes express five VAMPs in total. Vesicles with VAMP2, VAMP3, or VAMP5 were found near the ends of the myotubes. Some of these vesicles are also positive for caveolin-3, suggesting their participation in the development of T-tubules. Our morphological analyses revealed the characteristic arrangement of organelles in myotubes and the existence of transport vesicles near the ends of the myotubes.

Evaluation of ACL mid-substance cross-sectional area for reconstructed autograft selection.

PURPOSE: The purpose of this study was to compare the size of the native ACL mid-substance cross-sectional area and the size of commonly used autografts. Hypothesis of this study was that the reconstructed graft size with autografts would be smaller than the native ACL size. METHODS: Twelve non-paired human cadaver knees were used. The ACL was carefully dissected, and the mid-substance of the ACL was cross-sectioned parallel to the articular surface of the femoral posterior condyles at 90 degrees of knee flexion. The size of the cross-sectional area of the ACL, and the femoral and tibial footprints were measured using Image J software (National Institute of Health). The semitendinosus tendon (ST) and the gracilis (G) tendon were harvested and prepared for ACL grafts. Simulating an ST graft, the ST was cut in half. The bigger half was regarded as the antero-medial (AM) bundle, and the remaining half was regarded as the postero-lateral (PL) bundle. Simulating an ST-G graft, the bigger half of the ST and G were regarded as the AM bundle, and the smaller half of the ST was regarded as the PL bundle. Each graft diameter was measured, and the graft area was calculated. Simulating a rectangular bone-patella tendon-bone (BPTB) graft, a 10-mm-wide BPTB graft was harvested and the area calculated. RESULTS: The sizes of the ACL mid-substance cross-sectional area, femoral and tibial ACL footprint were 46.9 ± 18.3, 60.1 ± 16.9 and 123.5 ± 12.5 mm(2), respectively. The average areas of the ST, ST-G, and BPTB grafts were 52.0 ± 3.8, 64.4 ± 6.2, and 40.8 ± 6.7 mm(2), respectively. The ST and BPTB grafts showed no significant difference in graft size when compared with the ACL cross-sectional area. CONCLUSION: ST and BPTB autografts were able to reproduce the native size of the ACL mid-substance cross-sectional area. The ST-G graft was significantly larger than the ACL cross-sectional area. For clinical relevance, ST and BPTB grafts are recommended in order to reproduce the native size of the ACL in anatomical ACL reconstruction with autograft.

ACL footprint size is correlated with the height and area of the lateral wall of femoral intercondylar notch.

PURPOSE: The purpose of this study was to reveal the correlation between the size of the native anterior cruciate ligament (ACL) footprint and the size of the lateral wall of femoral intercondylar notch. METHODS: Eighteen non-paired human cadaver knees were used. All soft tissues around the knee were resected except the ACL. The ACL was cut in the middle, and the femoral bone was cut at the most proximal point of the femoral notch. The ACL was carefully dissected, and the periphery of the ACL insertion site was outlined on both the femoral and tibial sides. An accurate lateral view of the femoral condyle and the tibial plateau was photographed with a digital camera, and the images were downloaded to a personal computer. The size of the femoral and tibial ACL footprints, length of Blumensaat's line, and the height and area of the lateral wall of femoral intercondylar notch were measured with Image J software (National Institution of Health). RESULTS: The sizes of the native femoral and tibial ACL footprints were 84 ± 25.3 and 144.7 ± 35.9 mm(2), respectively. The length of Blumensaat's line and the height and area of the lateral wall of femoral intercondylar notch were 29.4 ± 2.8 mm, 17.1 ± 2.7 mm, and 392.4 ± 86 mm(2), respectively. Both the height and the area of the lateral wall of femoral intercondylar notch were significantly correlated with the size of the ACL footprint on both the femoral and tibial sides. CONCLUSION: For clinical relevance, the height and area of the lateral wall of femoral intercondylar notch can be a predictor of native ACL size prior to surgery. However, the length of Blumensaat's line showed no significant correlation with native ACL size.

The localization of VAMP5 in skeletal and cardiac muscle.

Vesicle-associated membrane protein 5 (VAMP5) is a member of the SNARE protein family, which is generally thought to regulate the docking and fusion of vesicles with their target membranes. This study investigated the expression and localization of the VAMP5 protein. Immunoblotting analyses detected the VAMP5 protein in skeletal muscle, heart, spleen, lung, liver, and kidney tissue, but not in brain or small intestine tissue. Through the immunofluorescence microscopy of skeletal muscle, we found that the expression level of VAMP5 varies among fibers. Most of the fibers with high expression levels of VAMP5 were categorized as type IIa fibers on the basis of their myosin heavy chain subtypes. In addition, the expression patterns of VAMP5 and glucose transporter 4 (GLUT4) were similar. In cardiac muscle, we determined that VAMP5 was localized to the vicinity of intercalated discs. These results suggest that VAMP5 plays local roles in membrane trafficking in skeletal and cardiac muscle.

Alterations of biochemical marker levels and myonuclear numbers in rat skeletal muscle after ischemia-reperfusion.

Prolonged ischemia-reperfusion results in various damages in skeletal muscle. Following reperfusion, although the damaged muscles undergo regeneration, the precise process and mechanism of regeneration have not yet been fully understood. Here, we show the altered levels of plasma biochemical markers of muscle damage, and the change in myonuclear numbers in adult rat skeletal muscle by ischemia-reperfusion. Male Wistar rats were subjected to unilateral hindlimb ischemia by clamping the anterior tibial artery for 2 h before reperfusion. Both plasma creatine kinase activity and C-reactive protein levels in plasma were increased significantly at 0.5 h of reperfusion and returned to the control level at 24 h. The transverse sectional area of muscle belly of the anterior tibial muscles in ischemic side was significantly decreased by 20 % compared with those in sham-ischemic (control) side at 2 days, and returned to the control level at 5 days of reperfusion. Moreover, the number of interstitial nuclei in the ischemic side were significantly increased at 5-14 days and returned to the control level at 21 days of reperfusion. Central nuclei that are specifically observed in regenerating muscle, appeared at 5 days, reached a peak at 14 days, and disappeared at 28 days of reperfusion. Furthermore, MyoD, a regulatory factor for myogenesis, showed a transient expression at 5 days of reperfusion. These results indicate that, although the size of muscle seems to be recovered by 5 days of reperfusion, the most active muscle regeneration occurs much later, as shown by the increase in central nuclei.

Hepatic CT perfusion measurements: a feasibility study for radiation dose reduction using new image reconstruction method.

OBJECTIVES: To assess the effects of image reconstruction method on hepatic CT perfusion (CTP) values using two CT protocols with different radiation doses. MATERIALS AND METHODS: Sixty patients underwent hepatic CTP and were randomly divided into two groups. Tube currents of 210 or 250 mA were used for the standard dose group and 120 or 140 mA for the low dose group. The higher currents were selected for large patients. Demographic features of the groups were compared. CT images were reconstructed by using filtered back projection (FBP), image filter (quantum de-noising, QDS), and adaptive iterative dose reduction (AIDR). Hepatic arterial and portal perfusion (HAP and HPP, ml/min/100ml) and arterial perfusion fraction (APF, %) were calculated using the dual-input maximum slope method. ROIs were placed on each hepatic segment. Perfusion and Hounsfield unit (HU) values, and image noises (standard deviations of HU value, SD) were measured and compared between the groups and among the methods. RESULTS: There were no significant differences in the demographic features of the groups, nor were there any significant differences in mean perfusion and HU values for either the groups or the image reconstruction methods. Mean SDs of each of the image reconstruction methods were significantly lower (p<0.0001) for the standard dose group than the low dose group, while mean SDs for AIDR were significantly lower than those for FBP for both groups (p=0.0006 and 0.013). Radiation dose reductions were approximately 45%. CONCLUSIONS: Image reconstruction method did not affect hepatic perfusion values calculated by dual-input maximum slope method with or without radiation dose reductions. AIDR significantly reduced images noises.

VAMP2 marks quiescent satellite cells and myotubes, but not activated myoblasts.

We examined the expression and intracellular localization of vesicle-associated membrane protein 2 (VAMP2) during the differentiation of skeletal muscle cells by immunofluorescence microscopy. In isolated single myofibers, VAMP2 was expressed in quiescent satellite cells, downregulated in proliferating myoblastic cells, and re-expressed with differentiation. In the myoblastic cell line C2C12, VAMP2 was expressed at a low level in the proliferating stage, and then increased after differentiation into myotubes. Based on these results, we propose that VAMP2 can be used as a molecular marker for both quiescent satellite cells and myotubes, but not for proliferating myoblasts. We also found the partial colocalization of VAMP2 with transferrin- or Rab11-labeled vesicles in myotubes, suggesting a role of VAMP2 in the trafficking of recycling endosomes.

Expression of protocadherin-19 in the nervous system of the embryonic zebrafish.

We have analyzed the expression pattern of protocadherin-19, a member of the delta2-protocadherins, in the nervous system of developing zebrafish using in situ hybridization methods. mRNA encoding protocadherin-19 (Pcdh19) began to be expressed at about 12 hours post fertilization (hpf) showing a segmental expression pattern in the anterior 1/3 of the neural keel, with strong expression in the presumptive forebrain, cerebellum/rhombomere 1 and rhombomere 4. Pcdh19 expression in the posterior neural keel was continuous and confined to the midline region. By 24 hpf, Pcdh19 was expressed widely in the brain and spinal cord, with higher expression levels in the ventral telencephalon, dorsal and central thalamus, optic tectum, central tegmentum, cerebellum and dorsolateral regions of the hindbrain. As development proceeded, Pcdh19 expression domains became restricted to the dorsal and/or lateral regions of the central nervous system, and Pcdh19 expression was not detected in the spinal cord of two- and three-day old embryos. Pcdh19 was also expressed by the eye primordium, developing retina, lens and otic vesicle. Similar to its expression in the brain, Pcdh19 expression in the eye and ear was also spatially and temporally regulated.

Molecular dissection of Otx1 functional domains in the zebrafish embryo.

Otx proteins are involved in the induction of neurectoderm patterning and morphogenetic movements, leading to the formation of the vertebrate central nervous system. Despite lack of homology of sequence outside the homeodomain, a large body of evidence has shown that the Otx/Otd class of proteins has similar functions in many animal phyla. Thus, characterization of functional domains in proteins of this family would help in understanding how this functional equivalence operates. Our previous analysis using the zebrafish embryo (Bellipanni et al., 2000, Dev Biol 223:339-353), has suggested that induction of cell aggregation is a morphoregulatory role of Otx/Otd factors in embryonic development. We now use the induction of cell aggregation as an in vivo assay to examine the functional requirement for particular domains of the zOtx1 protein. We demonstrate that zOtx1 induces cell aggregation by acting as a transcriptional activator through its C-terminal region. Further, we show that a region of 37 amino acids in the C-terminal third of zOtx1 is necessary but not sufficient for this activation potential. The effects of selective deletion of each of the three homeodomain alpha-helices of zOtx1 on cell aggregation were also tested. Surprisingly, we find that helix 3, which is required for binding to DNA, is dispensable for stimulation of cell aggregation. Our results suggest that for transcriptional activation of at least one gene in the cell aggregation pathway, zOtx1 need not bind directly to DNA, but does require helix 1 and 2 of its homeodomain to interact with an as yet undefined DNA binding protein.