The NH2-terminal amino acid sequence of cellular retinoic-acid binding protein from rat testis

Initial studies of the cytoplasmic FABP superfamily

Our colleagues and we have determined the complete primary structure of a low molecular weight cytoplasmic FABP (also known as z-protein) that binds to LCFAs with high affinities, obtained from rat liver. At the same time, we were the first to propose that rat FABP1, bovine FABP8 (MP-2), bovine CRBP and rat CRABP constituted a protein superfamily in 1982. Since then, extensive investigation of structures, functions and expressions has been carried out on a whole family of FABPs. Analyses of rat heart FABP; FABP1, FABP3 and alpha(2U)-globulin expressed in rat kidney; discovery of ileal FABP6 (I-15P); and first application of FABP2 as a diagnostic marker also stand out in particular.

Studies of the FABP family: a retrospective

Following my research on the role played by soluble proteins in their function as hydrophobic ligand carriers acting through squalene epoxidase, Dr Odani and I started to work together on low molecular lipid binding proteins. As a result of this collaboration, in 1982 we managed to determine the complete primary structure of Z-protein in rat liver. This was the first report ever to give the complete amino acid sequence of a fatty acid binding protein (FABP). This gave momentum to further such research, and now extensive exploration has been carried out on a whole family of homologous intracellular hydrophobic ligand binding proteins, the product of the expression of an ancient gene family in numerous organisms. Takahashi et al. have determined the primary structures of mammalian FABP family protein in liver, intestine, heart, kidney, and skin through amino acid sequencing as well as through determination of the cDNA sequence. Out of all my research on the FABP family, I believe, my initial study on FABP in liver, my work on kidney FABP, heart type FABP and my discovery of an I-15P (BAPB) and I-FABP application as a diagnostic marker stand out in particular.

Expression of CRABP-I and -II in human epidermal cells. Alteration of relative protein amounts is linked to the state of differentiation

The physiological role of cellular retinoic acid-binding proteins (CRABPs) may be to influence the intracellular level of free retinoic acid in the cell. In the present study two isoforms of CRABP, CRABP-I and CRABP-II were partially characterized in various human Malpighian epithelia and in human cultured keratinocytes expressing various patterns of differentiation. We have developed a new sensitive radiobinding assay using a PAGE/autoradioblotting technique which effectively separates CRABP-I and CRABP-II. This method allows the simultaneous quantification of these proteins. We show that CRABP-I and -II have similar M(r) values (15,000), but differ in their dissociation constant towards retinoic acid (Kd of 16.6 nM and 50 nM respectively), in pI (4.86 and 5.13) and in their relative mobilities (RF) on PAGE under nondenaturating conditions (RF values 0.65 and 0.44). In addition, we show that CRABP-II is the major isoform expressed in human keratinocytes, in vivo as in vitro. Furthermore, we demonstrate that CRABP-II is actually the CRABP previously studied in epidermal cells by a PAGE assay (Siegenthaler & Saurat (1987) Eur. J Biochem. 166, 209-214) and whose levels are dramatically increased by retinoic acid and its analogues in human epidermis. Keratinocytes, in the absence of full terminal differentiation, as well as hyperplasia, such as cultured human differentiating keratinocytes, psoriatic plaques, and non-keratinized oral mucosa, contained high levels of CRABP-II. CRABP-I was not detected in cultured keratinocytes, whereas normal skin (at full terminal differentiation) expressed CRABP-I and CRABP-II at a ratio of approx. 1:1.4. This value was approx. 1:17 in lesional psoriatic skin and 1:8 in oral mucosa. These observations suggest that CRABP-I and -II are regulated differently in human keratinocytes. The sharp increases in CRABP-II levels are associated with an alteration in the differentiation programme, as well as with cell response to retinoic acid overload, whereas CRABP-I might be a marker for terminal differentiation.

Intracellular actions of vitamin A

Because the effects of vitamin A vary with tissue type and often with the form of vitamin A itself, a complete understanding of the mechanism(s) of action still has not been attained. The action of vitamin A may be at the level of genomic expression, at the membrane level, or both. Intercellular and intracellular transport of vitamin A are facilitated by specific binding proteins but probably not in the cellular uptake of vitamin A. Subcellularly, vitamin A may exert a direct effect on transit through the Golgi apparatus, as observed from both biochemical and morphological studies. In my laboratory, recent work using cell-free systems has shown that retinol stimulates transition vesicle formation from endoplasmic reticulum in a GTP-requiring step.

Detection of cellular retinoic acid-binding protein in chick embryonic tissues by monoclonal antibodies

We describe the production of monoclonal antibodies (MAbs) directed against cellular retinoic acid-binding protein (CRABP) and their application for the quantitation and localization of CRABP during the development and growth of chick embryo. Three MAbs, classified as D-10 and H-6 (IgG1-isotype), and G-4 (IgM-isotype), exhibited the highest degree of immunoreactivity for chick embryo CRABP. The antibodies showed partial reactivity to CRABP from rat testis. None of the MAbs showed cross-reactivity with cellular retinol-binding protein or fatty acid-binding protein which are structurally similar to CRABP. The antigen-specificity was confirmed by immunoblot analysis as well as by fast protein liquid chromatographic analysis. The radioimmunoassay developed for MAb (D-10) provided a detectability range of 0.5-5.0 ng of CRABP in the standard displacement curves. An abundance of CRABP was found in embryonic skin, brain, testis and eye. Several other tissues (heart, lung, liver), previously reported to have undetectable levels of CRABP, showed significant amounts of the binding protein. The levels of CRABP peaked in early (4-6-day-old) and late (11-14-day-old) stages of embryo development. Immunolocalization of CRABP in chick embryo skin demonstrates a specific intense staining for the antigen in the dense areas of mesenchyme cells (mesodermal layers); little or no staining was apparent in the differentiated cells of epidermis and peridermis as well as in the loose connective tissues. The MAbs are useful not only in the purification of CRABP as an affinity adsorbent, but also in the elucidation of the possible role of CRABP in the transfer of the ligand to its nuclear receptors and in the morphogenetic gradient formation of RA in chick embryo tissues.

Immunohistochemical demonstration of liver fatty acid-binding protein in human hepatocellular malignancies

Twenty-three hepatoblastomas of childhood, sixty-two adult hepatocellular carcinomas, and two hepatic sarcomas were examined immunohistochemically with the use of a polyclonal antibody against rat liver fatty acid-binding protein (L-FABP), which cross-reacts to human L-FABP. All the hepatoblastomas and half of the hepatic cell carcinomas contained L-FABP immunoreactive tumour cells, whereas two hepatic sarcomas were negative. The overall frequency of immunostained tumour cells was 43.5 per cent in hepatoblastomas and 18.6 per cent in hepatocellular carcinomas, respectively. Histologically well-differentiated areas contained more numerous immunopositive cells than undifferentiated or immature ones. These results indicate that L-FABP immunoreactivity is a new candidate for a tumour cell marker in hepatic cell malignancies, although its biological role has not been elucidated.

Polyacrylamide gradient gel electrophoresis of cytosolic retinol- and retinoic acid-binding proteins: application to rat testis and liver

Distribution and cellular levels of retinol-binding protein and retinoic acid-binding protein, involved in the molecular action of retinoids, were analyzed in rat testis and liver. Both binding proteins of cytosolic extracts were separated by linear-polyacrylamide gradient gel electrophoresis and following electrophoretic separation, could be visualized by complementary identification tests such as autoradiography and marker proteins. The concentration of the binding proteins were evaluated by scanning the polyacrylamide gradient gels and the resulting data were found to be in accordance with those obtained by counting radioactivities. Polyacrylamide gradient gel electrophoresis appears suitable to detect and quantitatively evaluate cytosolic retinol- and retinoic acid-binding proteins.

A regulatory element within a gene of a ribosomal protein operon of Escherichia coli negatively controls expression by decreasing the translational efficiency

The trmD operon of Escherichia coli consists of the genes for the ribosomal protein (r-protein) S16, a 21 kDa protein (21K) of unknown function, the tRNA(m1G37)methyltransferase (TrmD), and r-protein L19, in this order. Previously we have shown that the steady-state amount of the two r-proteins exceeds that of the 21K and TrmD proteins 12- and 40-fold, respectively, and that this differential expression is solely explained by translational regulation. Here we have constructed translational gene fusions of the trmD operon and lacZ. The expression of a lacZ fusion containing the first 18 codons of the 21K protein gene is 15-fold higher than the expression of fusions containing 49 or 72 codons of the gene. This suggests that sequences between the 18th and the 49th codon may act as a negative element controlling the expression of the 21K protein gene. Evidence is presented which demonstrates that this regulation is achieved by reducing the efficiency of translation.

Cellular retinoic acid-binding protein and the role of retinoic acid in the development of the chick embryo

The distribution of cellular retinoic acid-binding protein (CRABP) in four stages of chick development is described using an affinity-purified antibody against rat CRABP. CRABP is the protein to which retinoic acid (RA) binds when it enters cells and may reflect the requirement of those cells for RA. We found several discrete cell populations which showed high levels of immunoreactivity. Some were in the neural tube such as the commissural neurons and the dorsal roof plate. Some were of neural crest origin such as the dorsal root ganglia, sensory axons, sympathetic ganglia, and enteric ganglia. The remaining populations were certain connective tissue cells, limb bud cells, and the myotome. These results suggest that certain organ systems, particularly the nervous system, have a requirement for RA during development and they may further our understanding of the teratogenic effects of retinoids on the embryo.

Cellular retinol-binding protein in mucosa, in benign and malignant non-squamous cell tumours of the head and neck

The measurement and localisation of cellular retinol-binding protein (CRBP), in samples of normal oral mucosa, larynx papilloma and malignant lymphoma of the oropharynx, was performed with a radioimmunoassay and immunolocalisation techniques. As compared with CRBP concentration in normal mucosa, those in laryngeal papilloma were significantly higher, but those of malignant lymphoma were similar. CRBP concentrations were highest in maturing keratinocytes within the prickle cell layers of normal mucosa and in laryngeal papillomas, as estimated on the basis of immunoreactivity to CRBP. The retinyl palmitate concentrations in extracts of oral mucosa correlated to the retinol concentrations, both in plasma and mucosal extracts, but not to the CRBP content in mucosal extracts. The immunolocalisation of a cellular retinoic acid-binding protein (CRABP) like antigen in normal oral mucosa showed much the same picture with strongest staining of the maturing keratinocytes of the prickle cell layers.

Purification of transfer RNA (m5U54)-methyltransferase from Escherichia coli. Association with RNA

tRNA (m5U54)-methyltransferase (EC 2.1.1.35) catalyzes the transfer of methyl groups from S-adenosyl-L-methionine to transfer ribonucleic acid (tRNA) and thereby forming 5-methyluridine (m5U, ribosylthymine) in position 54 of tRNA. This enzyme, which is involved in the biosynthesis of all tRNA chains in Escherichia coli, was purified 5800-fold. A hybrid plasmid carrying trmA, the structural gene for tRNA (m5U54)-methyltransferase was used to amplify genetically the production of this enzyme 40-fold. The purest fraction contained three polypeptides of 42 kDa, 41 kDa and 32 kDa and a heterogeneous 48-57-kDa RNA-protein complex. All the polypeptides seem to be related to the 42/41-kDa polypeptides previously identified as the tRNA (m5U54)-methyltransferase. RNA comprises about 50% (by mass) of the complex. The RNA seems not to be essential for the methylation activity, but may increase the activity of the enzyme. The amino acid composition is presented and the N-terminal sequence of the 42-kDa polypeptide was found to be: Met-Thr-Pro-Glu-His-Leu-Pro-Thr-Glu-Gln-Tyr-Glu-Ala-Gln-Leu-Ala-Glu-Lys- . The tRNA (m5U54)-methyltransferase has a pI of 4.7 and a pH optimum of 8.0. The enzyme does not require added cations but is stimulated by Mg2+. The apparent Km for tRNA and S-adenosyl-L-methionine are 80 nM and 17 microM, respectively.

Isolation and characterization of a cDNA clone corresponding to bovine cellular retinoic-acid-binding protein and chromosomal localization of the corresponding human gene

A bovine adrenal cDNA library was constructed and a clone corresponding to cellular retinoic-acid-binding protein (CRABP) mRNA was isolated and sequenced. The insert of the clone corresponds to 75 bp of the 5' untranslated portion, the whole translated and the complete 3' untranslated portion of the bovine CRABP mRNA. A genomic Southern blot, probed with CRABP cDNA, indicated that only one copy of the gene is present in the human genome. Hybridizing bands in restricted chicken and fish DNA were also observed. Using the CRABP cDNA as probe we have located the human CRABP gene to chromosome 3 in hybridizations to mouse-human, hamster-human and rat-human cell hybrids. In situ hybridizations on rat testis cells probed with CRABP and cellular retinol-binding protein antisense mRNA indicate that both proteins are expressed in tubuli cells.

Quantitation and tissue localization of the cellular retinoic acid-binding protein

The distribution of the cellular retinoic acid-binding protein (CRABP) in some rat tissues has been determined, and the protein has been localized by immunocytochemical techniques in sections from rat testis. In the testis CRABP was found in the seminiferous tubuli with Sertoli cells and the spermatogonia most intensely stained. All other cells of the germinal epithelium appeared largely devoid of CRABP. By use of an enzyme-linked immunosorbent assay CRABP was quantitatively estimated in several tissues and the highest levels were found in testis and eye. Comparisons of the tissue levels of CRABP and of the cellular retinol-binding protein (CRBP) did not reveal any apparent correlation.

Cellular retinoid binding proteins

The cellular retinol-binding protein (CRBP) and the cellular retinoic acid binding protein (CRABP) have similar physicochemical characteristics. The amino acid sequences of rat CRBP and bovine CRABP have been elucidated and they display 40% sequence identity. Both protein sequences appear to be evolutionarily highly conserved. The amino acid sequence of human CRBP, deduced from a cDNA-clone, is 96% identical to the rat CRBP sequence. CRBP and CRABP are members of a protein family, all members of which may bind hydrophobic ligands and interact with membrane components. All members of the protein family are probably related in tertiary structure and might interact with membrane components through two regions with a high probability for alpha-helix. The tissue distribution of CRBP and CRABP, together with their relation to lipid transporting proteins suggests that CRBP and CRABP are cellular transporting proteins for retinol and retinoic acid, respectively.

Developmentally regulated mRNAs in 3T3-adipocytes: analysis of transcriptional control

We have investigated the regulation of mRNA synthesis during 3T3-adipocyte differentiation by measuring the transcription of specific genes in isolated preadipocyte and adipocyte nuclei. Transcription was assayed by hybridization of newly synthesized RNA to cDNA clones coding for glycerophosphate dehydrogenase (GPD), the induced protein of 13K which is shown here to be related to myelin protein P-2, the induced protein of 28K, actin, and two RNAs that are not developmentally regulated. Transcription of GPD and 13K was observed in adipocyte but not preadipocyte nuclei. Actin was transcribed in both types of nuclei but at a lower level in adipocytes. For most of the RNAs examined, there was a consistent relationship between amounts of nuclear transcription and the abundance of the corresponding cytoplasmic mRNA in adipocytes. However, 13K and 28K mRNAs are 10-100 times more abundant than would be predicted by their nuclear transcription alone. Preliminary mRNA turnover experiments in which 5,6-dichloro-1-beta-D-ribofuranosylbenzimidazole was used to inhibit mRNA synthesis suggest that these mRNAs are much more stable in the adipocyte cytoplasm than the other mRNAs examined. These results indicate that the transcription of specific genes is increased during adipocyte differentiation and suggest that other levels of control, particularly mRNA stability, may contribute to the relative abundance of certain developmentally-regulated mRNAs in adipocytes.

Lipid binding activities of the P2 protein in peripheral nerve myelin

Lipid binding activities of the P2 protein in peripheral nerve myelin were examined using retinoic acid, retinol and oleic acid as ligands. The P2 protein showed the specific binding affinity to both of retinoic acid and retinol. The binding site of these ligands was suggested to be similar. In addition, the high binding activity of the P2 protein with oleic acid was also observed. The ligands specificities of the P2 protein are clearly different from those of cellular retinoic acid binding protein (CRABP), cellular retinol binding protein (CRBP), and Z protein. In amino terminal sequence, however, the P2 protein contained considerable homologous structure to these lipid binding proteins. Therefore, the P2 protein and these lipid binding proteins may belong to a family of structurally related proteins evolved from a common ancestral gene.

Amino-terminal amino acid sequence of the nonspecific phospholipid exchange protein from bovine liver

The amino-terminal amino acid sequence of the nonspecific phospholipid exchange protein from bovine liver has been determined. The first 52 amino-terminal residues in the sequence were identified. The sequence determined failed to show statistically significant homology to any previously published protein sequence. However, a stretch of 12 amino acids at the end of the sequence displays homology to the phosphatidylcholine-specific phospholipid exchange protein.

Identification and quantitation of intracellular retinol and retinoic acid binding proteins in cultured cells

Although the mechanism whereby vitamin A mediates normal cell differentiation and inhibits tumor cell proliferation is unknown, intracellular receptor-like proteins for retinol and retinoic acid have been implicated in the molecular action of vitamin A. We have assayed these two binding proteins, cellular retinol binding protein (protein R) and cellular retinoic acid binding protein (protein RA), in the cytosolic fraction of various normal and tumor cells via sucrose density gradient centrifugation and saturation analysis. Employing charcoal separation of bound and free tritiated retinoid, the saturation analysis yields an approximate Kd for ligand binding and an estimate of the number of protein R and protein RA molecules per cell. Unique protein R and protein RA macromolecules sedimenting at 2 S with Kd values of 7-42 nM are detected in murine cells (1 degree epidermal, 3T6 fibroblasts and melanoma) and human neuroblastoma cells. Concentrations of the intracellular binding proteins range from 55 000 to 3 000 000 copies per cell. When one cell line (C-127 mouse mammary) is transformed by bovine papilloma virus, protein RA levels increase from undetectable to 193 000 copies per cell. Assessment of growth inhibition by 10(-6) M retinol or retinoic acid in the culture medium reveals that there exists a partial, but not absolute, correlation between the presence of protein R or protein RA and the antiproliferative effect of the particular retinoid in the tested cell lines. We conclude that the 2 S intracellular binding proteins for the retinoids are present in most vitamin A responsive cells, but may not be essential for biologic actions of the vitamin such as growth inhibition in monolayer culture.

Molecular cloning of the temperature-inducible outer membrane protein 1 of Yersinia pseudotuberculosis

When Yersinia pseudotuberculosis is shifted from growth at 26 degrees C to growth at 37 degrees C, the synthesis of a plasmid-associated outer membrane protein, protein 1, is induced (Bölin et al., Infect. Immun. 37:506-512). The structural gene of this protein was found to be located on the virulence plasmid pIB1 of Y. pseudotuberculosis. One cosmid hybrid plasmid pBW8 was studied which carried a region of the virulence plasmid. This hybrid plasmid expressed in Escherichia coli K-12 a novel temperature-inducible outer membrane protein which is immunologically related to and has the same molecular weight as protein 1. Protein 1 was purified to homogeneity, and 14 amino acids of the N-terminal end were determined. From this sequence, the tentative corresponding DNA sequence was deduced, and a set of 11-nucleotide-long DNA probes was chemically synthesized. By using these probes in Southern blotting experiments, the genetic location of the N-terminal end of protein 1 was established. By introducing the transposon Tn5 into the virulence plasmid pIB1, mutants were obtained that did not express protein 1. One class of these mutants was still Ca2+ dependent and virulent, suggesting that protein 1 is not a major virulence determinant. Tn5-derived insertion mutants were also obtained which were Ca2+ independent. Such mutants were found to be avirulent. One Ca2+-independent mutant still expressed protein 1, indicating that the regulatory expression of protein 1 is not linked to Ca2+ dependence.

Isolation and characterization of the three fractions (DE-I, DE-II and DE-III) of rat-liver Z-protein and the complete primary structure of DE-II

Three fractions (DE-I, DE-II and DE-III) of Z-protein (fatty acid binding protein) have been isolated from rat liver cytosol by DEAE-cellulose chromatography and characterized. They had the same molecular weight of 14000 and essentially identical amino acid composition. However, compositions of endogenous fatty acids were found to differ strikingly from one another. Long-chain fatty acids detected in DE-II were palmitic, stearic, oleic, linoleic and arachidonic acids. In contrast to DE-II, DE-III contained mainly arachidonic acid. Molar ratios of endogenous long-chain fatty acids to both DE-II and DE-III were estimated to be around 1.0. Unlike the latter two fractions, DE-I was virtually lipid-free. Analyses of the three fractions by polyacrylamide gel electrophoresis, electrofocusing and DEAE-cellulose chromatography before and after delipidation suggested that the difference between DE-I and DE-II was in part due to fatty acids bound to DE-II. In contrast, DE-III appeared to be somewhat different from these forms in its protein structure, though tryptic peptide mappings of the three fractions did not reveal clear differences among them. Analysis of the primary structure was made on the most abundant fraction, DE-II, to investigate the relationship among the three fractions and to other proteins. The protein was a single chain consisting of 127 amino acid residues and had a mostly acetylated NH2 terminus and a free sulfhydryl group. The complete sequence of Z-protein showed striking homology to cellular retinoid binding proteins and peripheral nerve myelin P2 protein, which indicated the presence of a new family of cellular lipid-binding proteins diverged from a common ancestor. A possible intragenic duplication of Z-protein was also suggested.

Specific high-affinity binding and biologic action of retinoic acid in human neuroblastoma cell lines

Neuroblastoma cells are a good model for neuronal development because of their ability to extend neurites in response to various stimuli, including retinoic acid. In the present experiments, we have examined five human neuroblastoma cell lines (LA-N-1, IMR-32, LA-N-5, SK-N-MC, and CHP-100) for the presence of cellular retinoic acid binding protein (CRABP), a receptor-like protein implicated in the molecular functioning of vitamin A. CRABP is identified and quantitated by sucrose gradient centrifugation, selective inhibition by the mercurial reagent p-chloromercuribenzene sulfonic acid (PCMBS), and saturation analysis. All five lines contain significant levels of cytosolic CRABP (2.5-7.5 pmol/mg of protein), which display typical properties of specific high affinity retinoic acid binding, a sedimentation coefficient of 2 S, and inhibition by PCMBS. Three of the lines (LA-N-1, IMR-32, and LA-N-5) are strongly growth inhibited by 1 microM retinoic acid in monolayer culture, whereas two (LA-N-1 and LA-N-5) undergo marked differentiation to a stellate, fusiform morphology with characteristic neurite outgrowths. The SK-N-MC and CHP-100 lines are relatively resistant to the antiproliferative effects of retinoic acid under these conditions. Nevertheless, all five lines are effectively inhibited by retinoic acid in their ability to form anchorage-independent colonies in soft agar. Thus, although CRABP is not necessarily correlated with growth inhibition in monolayer culture, it is associated with retinoic acid's ability to inhibit neuroblastoma colony formation in soft agar. More experiments will be required to determine if this effect on growth in soft agar reflects the putative ability of retinoic acid to convert tumorigenic neuroblastoma cell lines into the normal differentiated phenotype.

Structure and tissue distribution of some retinoid-binding proteins

Vitamin A has, apart from its function in the visual pigments, general effects on several organs. Early signs of vitamin A deficiency include keratinization of epithelia and hyperkeratosis of the skin. To elucidate a generalized function for vitamin A, we have taken the approach of tracing the vitamin from its storage site in the liver via its blood transport by the retinol-binding protein (RBP) to its uptake by susceptible cells. We have also examined the intracellular occurrence of vitamin A as regards its binding to specific receptor proteins. Here we summarize data on the amino acid sequences of several vitamin A-binding proteins. The finding that CRBP and CRABP, the two intracellular proteins, are homologous to each other, to a myelin protein, and to a fatty acid-binding protein may shed light on the functions of these proteins. Retinoic acid, which binds to CRABP but not CRBP, induces differentiation of teratocarcinoma cells. This is accompanied by a lowering of the CRABP concentration, an increase of the CRBP level, and an increase in the uptake of retinol from RBP. The epidermis contains both CRBP and CRABP, and their distributions are rather similar. However, in contrast to CRBP, CRABP is most abundant in cells lining the hair follicles. CRBP occurs in greatest relative amounts in the outer layers of the epidermis. Since techniques have been developed to measure CRBP and CRABP, normal and disease-affected skin may now be explored as to quantity and cellular distribution of the retinoid-binding proteins.