Expression of the Lycat gene in the mouse cardiovascular and female reproductive systems

Authors

  • Weidong Wang,

    Corresponding author
    1. The Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine, Weill Medical College of Cornell University, New York, New York
    • The Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine, Weill Medical College of Cornell University, New York, NY 10065
    Search for more papers by this author
  • Lujuan Ni,

    1. The Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine, Weill Medical College of Cornell University, New York, New York
    Search for more papers by this author
  • Qingming Yu,

    1. Department of Medicine-Nephrology Division, Massachusetts General Hospital and Harvard Medical School, Charlestown, Massachusetts
    Search for more papers by this author
  • Jingwei Xiong,

    1. Department of Medicine-Nephrology Division, Massachusetts General Hospital and Harvard Medical School, Charlestown, Massachusetts
    Search for more papers by this author
  • Hung-Ching Liu,

    1. The Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine, Weill Medical College of Cornell University, New York, New York
    Search for more papers by this author
  • Zev Rosenwaks

    1. The Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine, Weill Medical College of Cornell University, New York, New York
    Search for more papers by this author

Abstract

The Lycat homologue in zebrafish maps to the deletion interval of the cloche mutant in which hematopoietic and endothelial cell lineages are affected. However, its definitive relationship to cloche is inconclusive, partly due to inadequate expression data of Lycat from any organisms. We precisely examined the temporal and spatial expression patterns of Lycat in mouse using RNA in situ hybridization, immunostaining, and BAC transgenesis. Lycat is initially expressed in developing heart, lung, and somites, and later becomes progressively restricted to all vascular smooth muscle cells. In adult ovaries, Lycat turns on in oocytes during the transition from primary to secondary follicles. Expression of the Lycat/reporter transgene in the extraembryonic mesoderm, cardiogenic mesoderm, and primitive streak, but not extraembryonic endoderm at E7.5, suggests its potential roles in regulating cardiac, smooth muscle, hematopoietic and endothelial lineages. Promoter mapping assay by transient transgenesis identifies a novel cardiac-specific regulatory region in the Lycat locus. Developmental Dynamics 239:1827–1837, 2010. © 2010 Wiley-Liss, Inc.

INTRODUCTION

We isolated mouse acyl-CoA:lysocardiolipin acyltransferase (Lycat) by subtractive screening between cDNAs from untreated and gonadotropin-primed mouse ovaries. Our ultimate goal of this project is to investigate the physiological function of this novel gene in controlling oocyte quality. Several other research groups are also working on different aspects of Lycat functions in lipid metabolism and hematopoietic stem cell determination (Cao et al.,2004; Wang et al.,2007; Xiong et al.,2008, Zhao et al.,2009).

Cardiolipin is a major membrane polyglycerophospholipid of the mitochondria, where it serves as a Ca2+-binding site in the inner mitochondrial membrane to regulate membrane permeability and potential (Schlame et al.,2000; Orrenius et al.,2003). Cardiolipin is the only known dimeric phospholipid, consisting of four fatty acyl chains, which is restricted to C18 chains dominated by the linoleoyl group (Schlame et al.,1993). The unique fatty acyl composition has been proven to be critical to the binding affinity of cardiolipin to proteins, thereby the biological functions. For example, dissociation of Cytochrome c from cardiolipin triggers apoptosis (Ott et al.,2002; McMillin and Dowhan,2002). However, the maturation of cardiolipin cannot be achieved by de novo biosynthesis (Rüstow et al.,1989). A newly synthesized cardiolipin precursor has to undergo a remodeling process to add appropriate acyl groups on it. The remodeling process is poorly understood because only a few enzymes potentially involved in this process have been characterized (Cao et al.,2004). Biochemical studies done by Shi's group demonstrate that recombinant Lycat does exhibit a preference to use linoleoyl-CoA and oleoyl-CoA as acyl donors. This finding is consistent with the unique fatty acid pattern of C18 found in mammalian cardiolipin (Cao et al.,2004; Hoch1992). Independent work done by Xiong's group maps the zebrafish lycat gene to the deletion interval of cloche, a mutant in which both endothelial and hematopoietic cell lineages are severely reduced (Stainier, et al.,1995; Xiong et al.,2008). Over-expression of Lycat during ES cell differentiation into embryoid bodies leads to enhanced formation of hematopoietic and endothelial cell populations as characterized by their expression of hematopoietic and endothelial marker genes such as Runx1, CD31, CD41, Flk1, Tie1, and VE-cadherin. Complementary to the gain-of-function assays in embryoid bodies, Lycat siRNA down-regulates a set of endothelial and hematopoietic genes (Wang et al.,2007). More interestingly, the mouse Lycat transgene was able to partially rescue the cloche mutant (Xiong et al.,2008). Despite the progress in characterizing Lycat as a cardiolipin-remodeling enzyme and its potential functions in hemangioblast development in vitro, little is known about its roles in the context of real physiological environments. In addition, its definitive relationship to cloche mutation is often questioned since there are no convincing expression data of Lycat in zebrafish. Even though the organ distribution of Lycat in mice has been examined by Northern blot analysis and RT-PCR (Cao et al.,2004; Wang et al.,2007), its temporal and spatial expression has not been carefully characterized, which is the first critical step to assess potential domains where Lycat exerts its functions. Given that most of the development control genes in the same gene family among different species are both structurally and functionally conserved, we assume that functional analyses of the mouse Lycat gene will provide instructive insights into its physiological roles in hematopoietic and endothelial cell fate determination as well as oocyte maturation in mice, thereby proving whether the Lycat gene is the long-sought cloche gene in zebrafish.

In the present study, we report the spatial and temporal expression patterns of the mouse Lycat gene at all developmental stages by in situ hybridization. We also employed a BAC transgenic approach to identify the regulatory elements involving the transcriptional control of the endogenous Lycat. Zhang and colleagues have made significant progress in chromosome engineering and they demonstrated that it is possible to modify DNA in Escherichia coli by homologous recombination using the RecE/RecT proteins of E. coli, named ET cloning (Zhang et al.,1998). The BAC transgenic approach has many advantages over conventional transgenes derived from plasmids. In general, BAC inserts are large enough to accommodate all regulatory elements of the gene of interest, and therefore confer accurate transgene expression in vivo. In addition, high stability and relatively low copy numbers of integrated BACs make positional effects and genome rearrangement less likely to happen. We modified two independent Lycat-containing BAC clones by inserting the ires.lacZ reporter gene into the 2nd exon of Lycat by ET-cloning. One BAC transgene, pW209, is able to faithfully recapitulate the expression patterns of the endogenous Lycat, indicating that all the regulatory elements of Lycat are well preserved in this BAC clone. Establishment of the pW209 BAC transgenic lines enables us to examine the dynamic expression profile of Lycat, identify the regulatory regions, and track the cell fate of Lycat-positive cells. Furthermore, using a transient transgenic approach, a novel cardiac-specific regulatory region is identified.

RESULTS

Expression Profile of Endogenous Lycat Examined by In Situ Hybridization

Figure 1 shows the expression patterns of mouse Lycat on paraffin sections of embryos examined by a digoxigenin or 35S-labeled antisense Lycat probe. When examined at embryonic day 9.5 (E9.5), Lycat has already been turned on in the developing heart (H), epithelial lining in the main bronchus within lung bud (L) and junctions between somites (Fig. 1A). Lycat expression in these organs persists throughout the embryonic stages. In the developing heart, Lycat is expressed in cells of both atria and ventricles (Fig. 1B,C). At E14.5 when the septation of each chamber is complete, a strong expression of Lycat can be seen in the mesenchymal cells (MCs) (morphologically, MCs are not tightly packed compared with neighboring cardiomyocytes) in the cardiac septa and valves (arrow in Fig. 1D), which originate from endocardial cells within the atrioventricular canal region through endothelial-to-mesenchymal transformation (EMT) (Zeisberg et al.2007; Niessen and Karsan,2008). Impaired EMT fails to generate precursor cells that populate the endocardial cushions, resulting in many congenital heart defects (reviewed by Markwald and Butcher,2007). In the developing lung, Lycat is strongly expressed in the epithelial cells of the main bronchus (Fig. 1A–E). Interestingly, the bronchial epithelial linings are of endodermal origin. In addition, notochord (NC in Fig. 1D) and cartilage primordium of the developing vertebrate are positive for Lycat (CP in Fig. 1D). In adult heart, Lycat is present in the entire atria (data not shown) and a subset of cells in the ventricles including the mesenchymal cells at the base of tricuspid valves (arrowheads in Fig. 1F). A small portion of cardiomyocytes underlying the endocardium in the ventricle also express Lycat (arrow in Fig. 1F). Cells in the smooth muscle layer of the thoracic aorta and the epithelial layer of trachea are also Lycat-positive (Fig. 1G,H). In the female reproductive organs, Lycat is strongly expressed in oocytes in the growing secondary follicles (Fig. 1I and J). Weak signals can be seen in the corpus luteum (data not shown).

Figure 1.

Lycat mRNA is expressed in the heart, aorta, lung and cartilage primordium, and maturing oocytes. Paraffin sections of wild-type mouse embryos and tissues were used for in situ hybridization with either digoxigenin-labeled or 35S-labeled full-length cDNA. Lycat signal was visualized by alkaline phosphatase staining (A-H; blue or purple signal) or autoradiography (I and J), respectively. Developmental stages are indicated at the bottom right of each panel. Lycat is expressed in the heart, and lung in sagittal sections of embryos at E9.5 (A), at E10.5 (B), and at E11.5 (C). A transverse section (D) and a sagittal section (E) show Lycat in the developing cardiomyocytes, lung epithelium, notochord (NC), and cartilage primordium of vertebrates (CP) at E14.5. Lycat signal is detectable in mesenchyme-like cells near valves and septum (arrowheads in F) in heart, smooth muscle cells in aorta (Ao in G), and epithelial cells in trachea (Tr in H) at 30 days postpartum. I, J: Bright-field view and the corresponding dark-field view of Lycat mRNA in a maturing oocyte after autoradiography. Ao, aorta; At, atrium; CP, cartilage primordium; dpp, day(s) after postpartum; H, heart; L, lung; NC, notochord; Tr, trachea; Vt, ventricle.

In summary, Lycat is specifically expressed in oocyte, heart, lung, the vascular structures associated to these organs, as well as the cartilage primordium of the vertebrate.

A 135-kb Genomic Fragment From the Lycat Locus Can Drive the Reporter Gene to Recapitulate the Endogenous Lycat Expression

The targeting strategy of this transgenic study is shown in Figure 2. The BAC clones RP23-294D3 and RP23-299H17 contain 170 and 135 kb of Lycat genomic DNA, respectively, but neither covers the entire Lycat gene (Fig. 2A). Mouse Lycat contains 6 exons that spread over 134 kb in length on chromosome 17. RP23-294D3 contains exons 2–6 and RP23-299H17 harbors exons 1–5 of Lycat. To map the Lycat regulatory elements, we first modified both BAC clones by inserting the reporter ires.lacZ into the second exon of Lycat by ET cloning to generate pW209 from RP23-299H7 and pW199 from RP23-294D3 (Fig. 2A). These transgenes were microinjected into one-cell embryos by pronuclear microinjection.

Figure 2.

Identification of the Lycat regulatory elements by ET cloning and transgenesis. A: The size and position of RP23-299H17 and RP23-294D3 Lycat BACs are indicated. Linearized targeting vector, pW207, was used to modify the BACs by ET-mediated homologous recombination in DH10B E. coli cells. Correct integration of the targeting vector results in the insertion of the reporter cassette into the 2nd exon of Lycat. Transgenic mice were generated by pronuclear injection of PI-SceI linearized BAC DNAs (pW199 and pW209). B: Erratic expression pattern of a pW199 transgenic embryo. a: A whole mount embryo stained for β-galactosidase activity; b: a sagittal section of the same embryo shown in a. Embryonic stage of the embryo is listed at the bottom right of each panel. C: Stable pW209 transgenic lines were verified by Southern blot analysis using a 32P-labeled neo probe. Approximately 5 μg of tail-tip genomic DNA from each stable transgenic line was digested with restriction enzyme EcoRI. The neo probe detected a 2.9-kb DNA fragment comprising the entire neo coding region and part of the Lycat BAC insert. Endogenous Umodl1 gene was included as the internal control. Copy number of pW209 in each transgenic line was estimated by TaqMan real-time PCR method (see Table in C).

First, transient transgenic assays were performed to examine whether the entire or part of the Lycat regulatory elements are included in these BAC clones. Embryos were collected from embryonic stage E8.5 to E12.5. Each embryo was genotyped by PCR using genomic DNA extracted from the attached yolk sac membranes. Expression of the Lycat BAC-ires.lacZ-A+ transgenes (Lycat:lacZ) was examined by β-galactosidase activity (lacZ-staining).

Totally, 13 pW199 transgenic embryos verified by PCR were collected for lacZ staining. These embryos showed no or erratic lacZ expression that is inconsistent with the endogenous Lycat (Fig. 2B). Lack of consistent expression of the pW199 transgenic embryos indicates that the 170-kb genomic fragment carried by the RP23-294D3 BAC clone contains no functional Lycat regulatory elements. However, when examined at E11.5, six out of eight pW209 transgenic embryos, as genotyped by PCR, showed identical expression patterns to those of the endogenous Lycat with minor variations in expression intensity, confirming that all of the Lycat regulatory elements are well preserved in RP23-299H17. Furthermore, absence of regulatory elements in pW199 suggests that the Lycat promoter/enhancers are primarily located in the 53-kb genomic fragment of RP23-299H17 that is not overlapped with RP23-294D3 (Fig. 2A).

Encouraged by these remarkable results from the above transient transgenic assays, we decided to establish stable pW209 transgenic lines so that we can use them to examine Lycat:lacZ expression patterns at any developmental stages. Additionally, these transgenic mice will serve as control animals for phenotypic analysis of both transgenic mice carrying extra copies of Lycat (over-expression) and the Lycat knockout mice in our future studies. After two rounds of pronuclear injection, 42 pups were born and 4 of them were transgenic as confirmed by PCR (data not shown) and Southern blot analysis (left panel in Fig. 2C). Transgene copy numbers were estimated by 2ΔΔCt TaqMan method using tail-tip genomic DNA samples from both the founders and their F3 offspring generated by crossing BAC transgene founder (F0) with FVB males/females (right panel in Fig. 2C). Absence of deviation in transgene copy number after several generations of breeding suggests a single site integration of the transgenic construct in all of these stable transgenic lines.

When their F1 embryos were tested by lacZ-staining, colonies pW209-17, 23, and 42 showed identical lacZ expression patterns to those of the endogenous Lycat. Transgenic line number 8 showed an irrelevant lacZ expression, which might be interpreted by positional effects or inter-genomic rearrangements of the BAC. Figures 3 and 4 show Lycat:lacZ expression of transgenic embryos and adult organs collected from Line pW209-17.

Figure 3.

Embryonic expression of the pW209 BAC transgene visualized by β-galactosidase activity. A–E: Whole-mount embryos showing dynamic expression profile of the BAC transgene from E7.5 (A) to E 11.5 (E). F–J: Paraffin sections of lacZ-stained embryos showing transgene expression in embryonic heart cells (F–I), epithelial cells of developing lung and midgut (G), and the vessel walls of umbilical artery (J). Sections were counter-stained with Nuclear Red. AER, apical ectodermal ridge; LA, left atrium; MG, midgut; RA, right atrium; So, somite; Tg, transgenic; UA, umbilical artery; UC, umbilical cord; WT, wild type. The rest of the abbreviations are as in Figure1.

Figure 4.

Expression of pW209 transgene in adult organs. A–E: lacZ-stained whole mounts of heart (B), lung (C), uterus (D), ovary (E), and associated vessels (A, D). F–T: Paraffin sections showing lacZ-positive cells. LacZ is detected in the trachea (A and F), pulmonary artery (Pa in A), aorta (Ao in A, K, and P), coronary vessels (arrow in B and arrowheads in Q), atrial myocardium (At in L), pericardium-derived mesenchymal cells at the base of the tricuspid valve (arrowheads in G and L), terminal bronchi or bronchioles of lung (C, H, and M), as well as a subset of nucleated blood cells (K and R). P: Higher-power view of the aortic smooth muscle layer. R: The same view of the boxed region in K at a higher magnification. In female reproductive organs, the transgene is seen in some of the maturing oocytes (E, J, and arrow in S) in the hemizygote ovary, as well as cells of smooth muscle lineage in the transgenic oviduct (S) and uterus (D, I, and N). In lactating mammary glands, the smooth muscle cells in the lactiferous ducts (arrowheads in O) and the epithelial cells in the extracellular matrix of the lactiferous lobes (arrows in O) are actively expressing the transgene. The lacZ-cells can also be found in bone marrow (T). Magnification of each section is indicated at the bottom right of each panel. AD, alveolar duct; AL, alveolus; Bn, bone; Bv, blood vessel; TriV, tricuspid valve; Pa, pulmonary artery. The rest of the abbreviations are as in previous figures.

In fact, Lycat:lacZ (pW209) transgene is detectable in all developmental stages examined from oocytes and fertilized eggs to adult mice; however, its expression domains are progressively restricted. At E7.5, Lycat:lacZ is expressed in the extraembryonic mesoderm (top arrow in Fig. 3F) and the cardiogenic plate (bottom arrow in Fig. 3F) and primitive streak, but not the extraembryonic endoderm (arrowheads in Fig. 3F). This supports the supposition that Lycat is expressed in the mesodermal precursors for generating hematopoietic and endothelial lineages, which is consistent with our previous findings in zebrafish and mouse embryonic stem cell–derived embryoid bodies (Wang et al.,2007; Xiong et al.,2008). At E8.5, no cells in the extraembryonic tissues including yolk sacs are positive for Lycat:lacZ (Fig. 3B). In addition to the heart, the lung bud, AER, and somites become positive for the transgene from E9.5 (Fig. 3C,D,E, and G). Interestingly, more cells in the atria are lacZ-positive than in the ventricles (Fig. 3G–I). β-galactosidase (lacZ) is also expressed in the epithelial layer of the third bronchial arch artery (arrow in Fig. 3H). At E11.5, lacZ is strongly expressed in the atrioventricular canal (arrow in Fig. 3I). These mesenchymal cells are critical for the development of septa and valves. The smooth muscle cells of umbilical artery are also Lycat-positive (Fig. 3J). When adult pW209 transgenic organs were examined, we found that Lycat:lacZ is restricted to the atrial myocardium (Fig. 4B and L), myocytes adjacent to ventricular endocardium (Fig. 4G and L), epicardium-derived mesenchymal cells at the base of the tricuspid valves and septa between chambers (arrows in Fig. 4G and L), and the smooth muscle cells of the coronary arteries (Fig. 4B and Q). In mature lungs, Lycat:lacZ is present in the periendothelial cells surrounding pulmonary arteries and veins (Fig. 4C,H, and M). In the arteries, veins, and airways associated with the heart or lung, LacZ is expressed in the epithelial layer of trachea and the smooth muscle layer of aorta (Fig. 4A,F,K, and P). Most strikingly, we found approximately 2% of the nucleated blood cells in the aorta and <1% of bone marrow cells are lacZ-positive (Fig. 4K, arrowhead in R and arrow in T). Previous studies demonstrate that Lycat is actively transcribed in hematopoietic and endothelial cell lineages, including the Flk1+ cells in embryoid bodies (EBs) derived from mouse ES cells and the Lin Sca+C-Kit+/CD31+ CD45 bone marrow cells (Wang et al.,2007). Whether these lacZ-positive cells are of hematopoietic and endothelial cell lineages remains to be investigated in the future.

In female reproductive organs, Lycat:lacZ is expressed in the smooth muscle layer of uterine blood vessels (Bv in Fig. 4I), myometrial cells (arrows in Fig. 4I and N), and the epithelial layer of uterine glands (arrowheads in Fig. 4I and N). At the embryo implantation sites, uterine expression of Lycat:lacZ is up-regulated in the myometrium and glandular epithelium (Fig. 4N). In the oviduct, Lycat:lacZ expression is only discernible in the smooth muscle layer (arrowhead in Fig. 4S). In the mature ovaries, approximately half of the growing secondary follicles are positive for the Lycat:lacZ transgene (arrows in Fig. 4E,J, and S), whereas the primordial follicles are Lycat:lacZ-negative (arrowheads in Fig. 4J). Sections of these ovaries indicate that the Lycat:lacZ signals are located in the cytoplasm of maturing oocytes (Fig. 4J and S). In the lactating mammary glands, the smooth muscle cells of the lactiferous ducts (arrowheads in Fig. 4O) and the epithelial cells in the extracellular matrix of the lactiferous lobes (arrows in Fig. 4O) are actively expressing Lycat:lacZ.

Immunofluorescence labeling was also performed to determine the identity of Lycat-expressing cells. Overlapped expression of Lycat-FITC and SM-MHC-TRITC reiterates the vascular smooth muscle lineage of Lycat-expressing cells (Fig. 5A and B). Furthermore, the pW209 transgene, which co-localizes with the endogenous Lycat, is excluded from the CD31+ vascular endothelial layer (Fig. 5C and D).

Figure 5.

Immunofluorescence micrographs of Lycat:lacZ transgenic (pW209) cardiac vessels labeled with the rabbit anti-mouse Lycat polyclonal (A), rabbit anti-bovine smooth muscle myosin heavy chain (B), and FITC-conjugated rat anti-mouse CD31 antibodies (C). Primary antibodies against Lycat and SM-MHC were visualized with FITC-conjugated goat anti-rabbit IgG and TRITC-conjugated goat anti-rabbit IgG secondary antibodies, respectively. Nuclei were stained with DAPI. D: Expression of Lycat:lacZ transgene in the vascular smooth muscle cells. Optical magnification is indicated at the bottom right of each panel.

In conclusion, our transgenic assay demonstrates that the RP23-299H17 BAC clone contains all of the Lycat regulatory elements, which are capable of driving the reporter gene to recapitulate the full spectrum of endogenous Lycat expression (Figs. 1, 3, 4, and 5).

Identification of a Cardiac-Specific Regulatory Region of Lycat

To identify the cis-regulatory element(s) responsible for cardiac/vascular smooth muscle/oocyte-specific expression of the Lycat gene, we first aligned the putative promoter regions of the human, mouse, and zebrafish Lycat genes, hoping that these cis-regulatory elements might have been well preserved during evolution. Unfortunately, no conserved binding motifs have been found among Lycat genes from different species. The Lycat genomic sequence was subjected for promoter scanning using databases including TransFAC (Heinemeyer et al.,1999) and ProScan (Bioinformation and Molecualr Analysis Section, NIH). Interestingly, as indicated in Figure 6A, the consensus binding motifs for various transcription factors are clustered in two genomic regions, named Cluster 1 and Cluster 2, which are located 30 and 38 kb from the 5′ end of RP23-299H17, respectively. To test the promoter activity of this proximal domain, a 9.2-kb genomic fragment containing Cluster 1, Cluster 2, and part of the putative first exon were amplified by PCR and fused to a reporter cassette, pW196a (Wang and Lufkin,2000). The resulting vector pW220 was used for transient transgenic assay (Fig. 5A). Figure 6B and C shows that this 9.2-kb proximal region of Lycat can only confer cardiac-specific transcription to the reporter. Embryonic expression of endogenous Lycat in somite, AER, lung bud, and bronchial arch artery was not observed, demonstrating that the 9.2 kb upstream sequence contains genetic information only for cardiac-specific expression, while other regulatory elements are likely located in the more 5′ distal region or the intragenic domain(s) of the Lycat gene.

Figure 6.

Mapping of the cardiac-specific promoter of mouse Lycat by transient transgenesis. A: Promoter scanning demonstrates that the cis-regulatory elements are clustered in two genomic regions at the 5′ end of Lycat. These consensus-binding motifs for various transcription factors are indicated. A 9.2-kb genomic fragment containing part of the 1st exon of Lycat was tested to drive reporter gene, ires.lacZ. After pronuclear injection, embryos were collected for promoter activity by lacZ-staining. B: This 9.2-kb genomic fragment is capable of directing the reporter gene specifically in the developing heart. The white line in B indicates the level and orientation of the section in C. C: A paraffin section shows cardiac-specific expression of the reporter gene, β-galactosidase. A, anterior; D. dorsal; P, posterior; and V, ventral. The rest of the abbreviations are as in previous figures.

DISCUSSION

Here, we employed in situ hybridization technique, immunofluorescence assay, and a BAC transgenic approach to map Lycat expression at cellular resolution. In short, the mouse Lycat gene is present at all developmental stages examined, including maturing oocytes, fertilized eggs, as well as adult circulatory and female reproductive organs. The data obtained from this study shows Lycat is first expressed in the primitive streak, extraembryonic mesoderm, and cardiogenic mesoderm at early embryonic stages. At the adult stage, aside from oocytes, Lycat is confined to the pericytes or vascular smooth muscle cells in all vascular structures, suggesting a potential role of Lycat in the development of blood vessels and the maintenance of vascular integrity. Expression in the extraembryonic mesoderm in E7.5 embryos enforces the assumption that Lycat is a master gene controlling hematopoietic and endothelial cell lineages in zebrafish and mouse embryoid bodies (Xiong et al.,2008; Wang et al.,2007). Lycat:lacZ-positive cells were indeed detected in the circulating blood and bone marrow. Whether these cells are Lin/Sca+/C-Kit+ hematopoietic stem cells (HSCs) is still obscure. We are currently attempting to isolate Lycat:lacZ-positive cells by FACS sorting fluorescein di-β-D-galactopyranoside (FDG)-stained cells from whole blood and bone marrow of the pW209 transgenic mice; hopefully, the molecular signature and pluripotency of the Lycat:lacZ-positive cells will be revealed soon.

Human and mouse Lycat proteins share as high as 89% homology in their amino acid sequences. Genbank BLAST search indicates that the Lycat genes belong to a novel acyltransferase family, which only shows a mild similarity (≈37%) to glycerol- phospholipidacyltransferase, a candidate gene responsible for Barth syndrome (Valianpour et al.,2002; Schlame et al.,2003). When expression of human LYCAT is examined by searching NCBI Expression and GenBank cDNA Support, a total of 124 hits was retrieved. Organs such as pancreatic islet (11), liver/spleen (11), breast (8), lung (6), and heart (3), display an abundant level of LYCAT. Surprisingly, LYCAT is also detected in many tumor tissues (23/124), including parathyroid tumor, hepatocellular carcinoma, chondrosarcoma, germ cell tumor, colon tumor, enchondroma, endometrial adenocarcinoma, and breast tumor. Elevated LYCAT expression likely results from high vascularization in these cancerous tissues.

Human LYCAT maps to chromosome 2p23 and its mouse homologue shows robust presence in the developing heart. Syndromic trisomy 2p congenital heart disease (CHD) involves a wide spectrum of defects in both heart and associated vascular structures, including inflow/outflow anomalies, malformed ventricles, and defective ventricular and atrial septa (Cassidy et al.,1977; Therkelsen et al.,1973; Lurie et al.,1995). Strikingly, more than 70% of all partial trisomy 2p cases published so far result from duplications of chromosomal band 2p23 (Lurie et al.,1995; Le Caignec et al.,2003; Aviram-Goldring et al.,2000; Willatt et al.,2001; Hahm et al.,1999), suggesting one or several genes on chromosome 2p23 are responsible for the patterning of organs including heart, craniofacial structures, and limb. Searching various databases including NCBI and Ensembl, we found approximately 117 different transcripts have been reported or predicted in this region, among which 71 genes are able to produce functional proteins. Interestingly, only three genes, FKBP1B (KBP12.6, FK506 binding protein 12.6), LBH (limb bud and heart homologue), and LYCAT, show a preferential expression in the heart. FKBP1B-deficient mice do show severe dilated cardiomyopathy, noncompaction of left ventricular myocardium, and ventricular septal defects that mimic a human congenital heart disorder. Additionally, some of the mutants exhibit exencephaly secondary to a defect in neural tube closure (Shou et al.,1998). Elegant work done by Alexandra Joyner's group has demonstrated that Lbh (limb bud and heart) gene is likely another candidate gene for CHD associated with partial trisomy 2p syndrome. Cardiomyocyte-specific over-expression of mouse Lbh resulted in phenotypes resembling human partial trisomy 2p syndrome, as manifested by impaired valvulogenesis of pulmonary outflow tract, abnormal cardiac septation, malformed inflow tract, as well as altered ventricular cardiomyocyte growth (Briegel et al.,2005). Expression data presented in the study strongly suggest Lycat may be another candidate gene responsible for the partial trisomy 2p syndrome.

Transgenic mice carrying modified bacterial artificial chromosomes (BACs) were initially used to map cell fates, ectopically express foreign genes, or introduce extra copies of genes of interest in the central nervous system (CNS) (Gong et al.,2003). In this study, we adopted the same strategy to track the lineage(s) of Lycat-expressing cells. The mouse Lycat gene is one of the largest genes in term of its genomic organization, whose 6 exons cover more than 134 kb on Chromosome 17E. Two independent Lycat BAC clones were engineered by inserting the reporter gene into the 2nd exon of the gene. Luckily, all the endogenous regulatory elements of Lycat are well preserved in one of these two clones, which can effectively specify reporter gene expression in cells that synthesize functional Lycat. This BAC/reporter transgenic assay is particularly advantageous for the in vivo study of Lycat function as it provides us with a clear fate map of Lycat-expressing cells. Our future effort will focus on dissecting out the minimal tissue-specific elements; subsequently, the transcription factors that bind to these elements will be identified. Regarding the Lycat signaling pathway, its downstream targets (or substrates) have been long sought. Despite the evidence that the prime target of Lycat is cardiolipins, we cannot rule out the possibility that Lycat is also involved in lipid modification of other signaling proteins essential for vasculogenesis and oocyte maturation. Post-translational modification including lipid modification is crucial for many signaling proteins to achieve their maximal activities. The best example is the dual lipid modification of Sonic Hedgehog (Shh) protein in regulating pattern formation during embryonic development and in maintenance of adult tissues. Shh is a secreted protein that controls cell proliferation, differentiation, survival, and migration (reviewed by Dessaud et al.,2008). Its initial posttranslational modification of Shh includes the signal peptide removal and auto-proteolysis of precursor protein into two fragments: a 27-kDa Shh-C and a 19-kDa Shh-N. The Shh-N fragment is further dually modified by adding a cholesterol moiety and a palmitate moiety onto the C- and N-terminal regions of Shh-N (Goetz et al.,2006). The modified Shh-N is the most active form of Shh that medicates long-range signaling in body patterning. It is likely that Lycat utilizes a similar strategy to exert its biological functions by modifying other signaling molecules.

Finally, our expression data suggest that, in addition to its original role in hematopoietic/endothelial cell lineage specification as in zebrafish, the mouse Lycat gene might have acquired novel functions in vascular smooth muscle cell determination during evolution. This study adds extra lines of evidence supporting the concept that the lycat gene is likely responsible for the cloche mutant phenotypes in zebrafish.

EXPERIMENTAL PROCEDURES

RNA In Situ Hybridization

Embryos between the ages of embryonic day (E) 9.5 and E15.5, and adult organs including hearts, lungs, and ovaries, were collected and fixed in 4% paraformaldehyde overnight. The entire coding region of the Lycat cDNA was amplified from the mouse ovarian total RNA by RT-PCR using Oligo 143 (5′-GCTTCGGGATGAATTAGCGGCG GGTTCT-3′) and Oligo 144 (5′CATTCCCGTGACAACCCTAGGTTTTCACATG AAC-3′). This 1.3-kb Lycat cDNA was then subcloned into pGEM-T (Promega, Madison, WI) resulting in plasmid pW52a. The Lycat in situ probe was synthesized using Sp6 RNA polymerase (New England Biolabs, Ipswich, MA) from the HindIII-linearized pW52a. Ovarian sections were hybridized with a 35S-labeled Lycat RNA probe. Specimen treatment, hybridization, washing, and autography were essentially performed as previously described (Wang et al.,1998; Wang and Lufkin,2000). Developed sections were finally counter-stained with haematoxylin. Other sections presented in this report were hybridized with the digoxigenin-labeled Lycat RNA probe as proposed by Jin and Lloyd with some modifications (Jin and Lloyd,1997). Briefly, DIG RNA labeling Kit (Cat. no. 11175025910) from Roche (Nutley, NJ) was used to prepare the antisense Lycat probe. The Dig-labeled RNA fragment was hydrolyzed in Hydrolysis Solution (80 mM NaHCO3, 120 mM Na2CO3, and 10 mM DTT) for 60 min and precipitated for subsequent hybridization. Paraffin sections were first rehydrated and primed with Proteinase K, followed by dehydration through graded ethanols. Pre-treated sections were covered with 150 μl Hybridization Buffer (50% formamide, 10% dextran sulfate, 1× Denhart's solution, 10 mM Tris-HCl, pH 7.5, 600 mM NaCl, 1 mM EDTA, 0.25% SDS, and 0.5 mg/ml yeast tRNA) containing 100 ng/ml of DIG-labeled Lycat RNA probe, and hybridized at 55°C overnight. After hybridization, slides were washed twice with 2× SSC, followed by 0.2× SSC once at 45°C to remove unbound probe. The sections were then incubated with Blocking Buffer (0.1M Tris-HCl, pH 7.5; 0.15M NaCl; saturated with block reagent) for 1 hr at room temperature. Alkaline phosphatase-conjugated anti-DIG antibody (Anti-DIG Fab AP; Roche, Cat. no. 11093274910) was diluted by 1:150 in Blocking Buffer and applied to each slide. After a 2-hr incubation at room temperature, sections were rinsed three times with Detection Buffer (0.1M Tris-HCl, pH 9.5, 0.1M NaCl, and 50 mM MgCl2), and then covered with Detection Buffer containing 0.18 mg/ml BCIP and 0.34 mg/ml NBT. This chromogenic reaction was carried out at 4°C for 16 hr. Finally, all slides were washed with 1× TE buffer, mounted with coverslips, and photographed.

Transgene Construction and Transgenic Mice Generation Modification of Lycat BAC Clones

The Lycat BAC clones, RP23-294D3 and RP23-299H17, were purchased from the Children's Hospital Oakland Research Institute. To generate the targeting vector to modify Lycat BAC clones, a 492-bp genomic fragment spanning the 2nd exon of Lycat was PCR amplified using Oligo 472 (5′-TTCCAGACAGCAGAGTTTCCTGCA-3′) and Oligo 473 (5′-CTCCTGAGAACTAGACCTCCGATGCT-3′), and subcloned into the SmaI site of pBSSKII(−), generating plasmid pW202a. pW202a was then digested with EcoRV and SalI, end-blunted with Klenow, and self-ligated to eliminate the HindIII site in the polylinker, further resulting in a construct pW203. The reporter cassette, ires.lacZ/GT1.2-neo, was inserted into the unique HindIII site located in the 2nd exon of Lycat in pW203, finalizing the targeting vector pW207. The 5′ and 3′ targeting arms are 321 and 171 bp, respectively. Preparation of recombinogenic-competent E. coli cells, transformation of pRedET and targeting fragment, and L-arabinose induction were performed essentially as described by Dr. Francis Stewart's laboratory (http://www-db.emblheidelberg.de/jss/servlet/de.embl.bk and http://wwwTools.GroupLeftEMBL/ExternalInfo/stewart/ETprotocols.html). Integration of the targeting vector into the BAC clones by homologous recombination was confirmed by PCR amplification of the region around the insertion sites using Oligo 292 (5′-GCATCGCCTTCTATCGC-3′) and Oligo 475 (5′-CAATCTGTTCGCTTCATCT-3′), and then further confirmed by sequencing the integration junctions. The integrity of the modified BAC clones was verified by comparing the digest pattern of the recombinant BACs to that of the original BAC DNA (data not shown). Modified RP23-294D3 and RP23-299H17 were renamed as pW199 and pW209, respectively (Fig. 2A). BAC DNAs were isolated by the alkaline lysis method (Qiagen Kit, Cat. no. 12162) and linearized by PI-SceI digest.

Transgene Construction for Identifying the Proximal Regulatory Elements of Lycat

A 9.2-kb genomic fragment containing part of the Lycat exon 1 was amplified by PCR using Oligo 353(5′-GCACAGCCTCCAATACACTGCTGCTGTGAGTTCC-3′) and Oligo 354 (5′-GGGACACACCGGGGGAAACGCGCG-3′). This putative promoter/enhancer region was blunted with Klenow and subcloned into the EcoRV site of pBSSK(−)II, resulting in plasmid pW137. Transgene pW220 is derived from pW137 by inserting the ires.lacZ/GT1.2-neo cassette into the unique XhoI site (Fig. 6A). Orientation of the ires.lacZ/GT1.2-neo was determined by restriction enzyme digestion and DNA sequencing. The Lycat-lacZ (pW220) insert was released from the vector by NotI and SalI double digestion.

Injection of the modified BAC constructs, as well as the Lycat-lacZ plasmid construct into one-cell FVB embryos was performed in our laboratory. Generation of transient and stable transgenics, Southern blot analysis, and β-galactosidase staining were performed as previously described (Wang and Lufkin,2000; Wang et al.,2001). β-galactosidase-stained embryos or tissues were embedded in paraffin and sectioned. Wild type embryos or adult tissues were also collected as controls for β-galactosidase staining in this study to ensure that the observed β-galactosidase activity results from the transgene expression.

TaqMan Real-Time PCR Method to Estimate Copy Numbers of Transgene

TaqMan copy number determination assay, essentially the comparative CT (2ΔΔCt) method, was performed as suggested by Applied Biosystems (Foster City, CA) (Livak et al.,1995; Schmittgen and Livak,2008). The threshold cycle value difference (ΔCt) between NeoCt and ApoBCt of each reaction was compared to the Ct value of diploid ApoB. The copy number of pW209 in each transgenic line was calculated to be two times the relative quantity. Genomic DNA was extracted from the tail-tips of F0 and F3 mice of all four independent pW209 transgenic lines (Tg8, 17, 23, and 42). Reference and target primers and probes are: (A) Internal control: mouse apolipoprotein B (ApoB), forward primer 5′TGGCAAACACTTACGGGTC-3′; reverse primer: 5′-TGGCTGTTAGAATG CTGGAG-3′; probe: 5′/VIC/CAAAAGTTGAATCTCAGCACGTGGGC/TAMRA/-3′.

(B) Transgene specific Neo: forward primer 5′-CAAGGTGAGATGACAGGAGATC-3′; reverse primer: 5′-TGAACTGCAAGACGAGGC-3′; probe: 5′-/FAM/CACTGAA GCGGGAAGGGACTGG/3lABkFQ/-3′.

All samples were run in triplicate in a 20-μl reaction volume containing 2× TaqMan Universal PCR Master Mix (4304437; Roche, NJ), primers/probes and 20 ng of genomic DNA. The PCR was run in the 7900HT Real-time PCR system (Applied Biosystems) using the following amplification parameters: 10 min at 95°C, and 40 cycles of 15 sec at 95°C and 1 min at 60°C. Applied Biosystems CopyCaller™ Software was used to calculate the transgene copy numbers by relative quantitation.

Immunofluorescence Labeling Assay

A pW209 transgenic mouse heart was collected, embedded into Neg-50 (Richard-Allan Scientific, Waltham, MA), and stored at −80°C for 20 min. Tissue was then cryo-sectioned at 7 μm. Adjacent slides were selected and incubated at 50°C for 2 hr before antibody labeling or lacZ-staining.

Immunolabeling was performed with a rabbit anti-mouse Lycat polyclonal antibody, a rabbit anti-bovine SM-MHC polyclonal antibody (ab53219, Abcam Inc., Cambridge, MA) and a FITC-conjugated rat anti-mouse CD31(PECAM-1, MEC13.3) antibody (553372; BD Biosciences, San Jose, CA) on adjacent cryo-sections. The rabbit anti-mouse Lycat polyclonal antibody was generated by Biosource Incorporation, MA. Lycat and SM-MHC primary antibodies were visualized by staining with FITC-conjugated goat anti-rabbit (F9887; Sigma, St. Louis, MO) and TRITC-conjugated goat anti-mouse secondary antibodies (sc-3841; Santa Cruz Biotechnology, Inc., Santa Cruz, CA), respectively. To suppress nonspecific binding, cryosections were pre-incubated with Blocking Solution (1×PBS containing 2% normal goat serum, 0.2% Triton X-100, 2% BSA, and 1% DMSO) for 1 hr at room temperature. Primary antibodies were diluted by 200-fold in the Blocking Solution. Incubation with primary antibodies was performed for 16 hr at 4°C. Secondary antibodies were diluted in Washing Buffer (1× PBS with 0.05% Tween-20) by 400-fold. Incubation with secondary antibodies was performed for 1 hr at room temperature. Slides were then washed with Washing Buffer, followed by DAPI (sc-24941; Santa Cruz Biotechnology, Inc.) staining. Tissue sections were evaluated and photographed with a Nikon (Melville, NY) Eclipse 90i fluorescence microscope with appropriate filter systems. Meanwhile, a comparable section was subjected to lacZ-staining to confirm that the pW209 transgene co-localizes with the endogenous Lycat protein.

Acknowledgements

We thank Mr. Zhiying He for photographing the fluorescent images. W.W. and Z.R. conceived this project, interpreted the results, and wrote the article. W.W designed the experiments, performed the vector construction and microinjection. L.N. maintained the mouse colonies, and performed most of the histological work. Q.Y. and J.X. raised the rabbit anti-mouse Lycat polyclonal antibody. We are grateful to the Ronald O. Perelman and Claudia Cohen Center for Reproductive Medicine, Weill Medical College of Cornell University for funding (to Z.R).

Ancillary