Smoothing random noise from human head scan data

In obtaining human head range data with a laser scanner, noise results from system errors and rough surfaces. The noise can be reduced with a suitable filter. Two aspects need to be considered in smoothing head scan data: one is finding a filter to eliminate noise without creating new artifacts; the other is determining the parameters of this filter to achieve optimal smoothing. The Gaussian filter has been shown to have unique characteristics which preserve the integrity of the data. A cross-validation method based on regularization theory has been derived for estimating the correct filter size for smoothing head range data. The authors discuss the justification and implementation of the statements above. Generalized cross-validation is derived in the two-dimensional case. Experimental results are presented that show the technique is effective and robust.

Processing of the envelope glycoprotein gp160 in immunotoxin-resistant cell lines chronically infected with human immunodeficiency virus type 1

We describe the isolation and characterization of variant cell lines which are chronically infected with the human immunodeficiency virus (HIV) and resistant to the action of immunotoxins directed against the HIV envelope protein. These variants all produce normal levels of HIV proteins, budding virions, and the envelope protein precursor gp160. Two of the variants, 10E and 11E, contain a mutation within the env gene which results in the production of a truncated precursor and altered processing and transport of the protein to the cell surface. Variants B9 and G4 are defective in gp160 cleavage and do not efficiently transport the envelope protein to the cell surface. There are no mutations in the expressed viruses of B9 and G4. These cell lines express higher levels of CD4 protein and mRNA than H9/NL4-3. Thus, 10E, 11E, B9, and G4 have escaped immunotoxin action by downmodulating the envelope protein from their cell surfaces. None of these variants produce infectious HIV. Two other immunotoxin-resistant variants, E9-3 and 41-17, produce normal levels of gp160, efficiently transport the cleaved and processed subunits to the cell surface, and secrete infectious HIV. These studies identify alterations in gp160 processing that underscore the importance of the relationship between HIV and the cell that it infects.

A mutation affecting signal peptidase inhibits degradation of an abnormal membrane protein in Saccharomyces cerevisiae

Signal peptidase removes amino-terminal signal peptides from precursor proteins during or immediately following their translocation to the lumen of the endoplasmic reticulum (ER) and may participate in ER degradation, a poorly defined process whereby abnormal proteins are rapidly degraded early in the secretory pathway. Here, the involvement of signal peptidase in ER degradation is examined through the use of two chimeric membrane proteins that lack amino-terminal signal peptides: A189invHD, which contains sequences derived from arginine permease and histidinol dehydrogenase, and AHDK2, containing the ER-resident protein Kar2p fused to the carboxyl terminus of A189invHD. Degradation of approximately 95% of A189invHD is observed in yeast cells expressing enzymatically active signal peptidase, whereas only 60% undergoes rapid degradation in a sec11 mutant bearing a temperature-sensitive mutation in the gene encoding the 18-kDa subunit (Sec11p) of the signal peptidase complex. AHDK2 is proteolyzed in a reaction yielding at least two fragments in wild-type cells and in the sec11 mutant containing a plasmid bearing the SEC11 gene. The proteolytic reaction is catalyzed in a temperature-dependent manner in the sec11 mutant, with AHDK2 remaining stable at the nonpermissive temperature. Using conditional mutants defective in protein translocation into and out of the ER and in vitro protease protection studies, the site of degradation for AHDK2 is localized to the ER lumen. The data therefore indicate (i) A189invHD is degraded through both signal peptidase-dependent and independent processes; (ii) signal peptidase, specifically the Sec11p subunit, is required for the proteolysis of AHDK2; and (iii) the Kar2 fragment at the carboxyl terminus of AHDK2 permits detection of proteolytic intermediates.

Trehazolin, a slow, tight-binding inhibitor of silkworm trehalase

Mechanisms of enzyme inhibition by trehazolin, a new inhibitor of trehalase (Ando et al. (1991) J. Antibiot. 44, 1165), were investigated using purified soluble silkworm trehalase and other glycosidases. Trehazolin inhibited trehalase with an IC50 value of 27 nM, whereas some other exo-alpha-glucosidases were inhibited only weakly, with IC50 values ranging from 7 to 370 microM. Other glycosidases tested were not inhibited by 500 microM trehazolin. The inhibition of trehalase by trehazolin was competitive with respect to trehalose. A notable feature of the inhibition was a slow progression of the association and dissociation of the enzyme-inhibitor complex. Preincubation of the enzyme and the inhibitor at 37 degrees C potentiated the inhibition by 10-times in a time-dependent manner up to 6 h. Dialysis of the inactivated enzyme recovered the enzymatic activity very slowly, and the rate constant for the dissociation at 37 degrees C was (7.3).10(-2) h-1. Trehalamine, a deglucosylated form of trehazolin, inhibited both silkworm trehalase and exo-alpha-glucosidases only weakly. The inhibition of trehalase by trehalamine was reversible. Rat isomaltase inhibition by trehazolin and sucrase inhibition by trehalamine were also reversible. Taken together, trehazolin is a specific slow, tight-binding inhibitor of trehalase, and the glucose moiety of the inhibitor is essential to the tight binding.

Numerical modeling of short-pulse excimer lasers with negative branch unstable cavities

A one-dimensional code for the numerical simulation of negative branch unstable resonators with an intracavity aperture that are applied to high-gain, short-pulse XeCl lasers is described. The model predicts near- and far-field performance of the output laser beams. The intracavity aperture size is shown as an important parameter for control of the output beam energy and divergence. A comparison with experimental measurements is presented.

[A comparison analysis on methods of astigmatism examinations for children]

70 children 140 eyes with astigmatic refractive error were reported. The methods of examination included retinoscopy, autorefraction and keratometry. The data of eye astigmatism and keratic astigmatism measured by the methods were analyzed. It revealed that there existed a good identity among the three methods, and the relationship between eye astigmatism (EAs) and keratic astigmatism (KAs) shown by the formula EAs = 0.94 KAs - 0.13 was similar to that of previous reports. It is suggested that the keratometry be necessary as an assistant method of refraction to assure to get a reliable astigmatism result for children.

Myeloglycan, a series of E-selectin-binding polylactosaminolipids found in normal human leukocytes and myelocytic leukemia HL60 cells

Sialosyl-Lex (SLex) is assumed to be the binding epitope of E- and P-selectin in normal human neutrophils and myelocytic leukemia HL60 cells. Glycosphingolipid (GSL) fractions from large quantities of normal human neutrophils and HL60 cell extract did not contain SLex GSLs having 6-10 sugar residues, as commonly found in solid tumor cells and tissues. Instead, the binding target of E-selectin was revealed to be a series of long-chain, unbranched polylactosamine GSLs with terminally sialylated, internally alpha 1-->3 polyfucosylated structure as the major component, or having SLex at the terminus and internally polyfucosylated structure as a minor component. These GSLs are hereby collectively termed "myeloglycan." Regardless of the site of fucosylation, all myeloglycans cross-react strongly with "anti-SLex" monoclonal antibodies such as CSLEX, FH6, SNH3, and SNH4.

Unique insertion sequence and pattern of CD4 expression in variants selected with immunotoxins from human immunodeficiency virus type 1-infected T cells

To study the variability of human immunodeficiency virus type 1 (HIV-1), we used immunotoxins to select for variants within a population of H9 cells persistently infected with a molecular clone of HIV-1 designated NL4-3. Chimeric immunotoxin CD4-PE40 (a chimeric fusion protein consisting of the amino-terminal two domains of CD4 and the carboxy-terminal domains of Pseudomonas exotoxin A) was used to select for cells lacking cell surface expression of HIV Env (envelope proteins gp160, gp120, and gp41). The cells described here (A1, A7, C9, and E9) fail to express HIV proteins because they have markedly diminished transcription of the integrated provirus (A1, A7, and E9) or no HIV provirus (C9). Analysis demonstrated that two different cloned variants, A1 and E9, contain the complementary sequence of tRNA(3Lys) (45 bp) inserted 3' to the primer-binding site, following by a 169-bp deletion through the start of the gag gene. No HIV mRNA was detected by Northern (RNA) blotting, but PCR demonstrated the presence of the viral message. These variants were found very infrequently in the unselected H9/NL4-3 cell population, and they contained proviruses distinct from that found in the dominant population. In addition, all of these variants had similar patterns of CD4 surface expression that allowed them to escape reinfection within the tissue culture. The data are discussed with regard to mechanisms and errors of HIV reverse transcription, as well as the evolution of mutants within a population of persistently infected cells.

Evolution of actin gene families of sea urchins

The actin gene family of the sea urchin Lytechinus pictus includes a single muscle actin gene, LpM, and four cytoskeletal actin genes: LpC1, LpC2, LpC3, and LpC4. The origin and relationship of these actin genes to members of the actin gene family of the sea urchin Strongylocentrotus purpuratus were considered. Comparison of deduced amino acid sequences suggested a close relationship between LpC1 and the CyI-CyII subfamily of S. purpuratus actin genes, and between LpC2 and the CyIII subfamily of S. purpuratus actin genes; the muscle actin genes were orthologous. It is proposed that two divergent cytoskeletal actin genes of the common ancestral sea urchin gave rise by duplication to the extant cytoskeletal actin genes of these species, some of which have changed 3' noncoding sequences while others have maintained a terminus highly conserved among sea urchin actin genes.

Nonlethal sec71-1 and sec72-1 mutations eliminate proteins associated with the Sec63p-BiP complex from S. cerevisiae

The sec71-1 and sec72-1 mutations were identified by a genetic assay that monitored membrane protein integration into the endoplasmic reticulum (ER) membrane of the yeast Saccharomyces cerevisiae. The mutations inhibited integration of various chimeric membrane proteins and translocation of a subset of water soluble proteins. In this paper we show that SEC71 encodes the 31.5-kDa transmembrane glycoprotein (p31.5) and SEC72 encodes the 23-kDa protein (p23) of the Sec63p-BiP complex. SEC71 is therefore identical to SEC66 (HSS1), which was previously shown to encode p31.5. DNA sequence analyses reveal that sec71-1 cells contain a nonsense mutation that removes approximately two-thirds of the cytoplasmic C-terminal domain of p31.5. The sec72-1 mutation shifts the reading frame of the gene encoding p23. Unexpectedly, the sec71-1 mutant lacks p31.5 and p23. Neither mutation is lethal, although sec71-1 cells exhibit a growth defect at 37 degrees C. These results show that p31.5 and p23 are important for the trafficking of a subset of proteins to the ER membrane.

Functional role of N-glycosylation in alpha 5 beta 1 integrin receptor. De-N-glycosylation induces dissociation or altered association of alpha 5 and beta 1 subunits and concomitant loss of fibronectin binding activity

Fibronectin (FN)-mediated cell adhesion is controlled mainly by alpha 5 beta 1 (recognizing the RGD sequence) and alpha 4 beta 1 (recognizing the CS-1 peptide sequence of FN) integrin receptors. Integrin-dependent cell adhesion to FN is greatly promoted by optimal GM3 concentration at the surface membrane (Zheng, M., Fang, H., Tsuruoka, T., Tsuji, T., Sasaki, T., and Hakomori, S. (1993) J. Biol. Chem. 268, 2217-2222), and cell adhesion mediated by alpha 4 beta 1 (to FN) or alpha 6 beta 1 (to laminin) is inhibited by modifying N-glycosylation processing of the integrin receptor (e.g. Akiyama, S. K., Yamada, S. S., and Yamada, K. M. (1989) J. Biol. Chem. 264, 18011-18018). We therefore studied the specific role of N-glycosylation in alpha 5 beta 1 function. Key findings of the present study were as follows. (i) Adhesion of K562 cells to FN-coated plates, which is mediated solely by alpha 5 beta 1, was inhibited when cells were treated with a mixture of endo-N-acetylglucosaminidase F and peptide -N4-(N-acetylglucosaminyl)asparagine amidase F (endo-F/PNGase-F). (ii) The alpha 5 beta 1 receptor at the K562 cell surface tended to dissociate into alpha 5 and beta 1 subunits when an extract of cells treated with endo-F/PNGase-F was precipitated by integrin subunit-specific antibodies, i.e. the alpha 5 subunit was preferentially precipitated by anti-alpha 5 monoclonal antibody ZH5, and the beta 1 subunit was preferentially precipitated by anti-beta 1 monoclonal antibody ZH1. When intact cells were extracted and treated with either ZH5 or ZH1, both alpha 5 and beta 1 were coprecipitated, indicating that the two subunits are normally tightly associated with each other. (iii) Adhesion of alpha 5 beta 1-containing liposomes (phosphatidylcholine:cholesterol liposomes incorporating purified alpha 5 beta 1) to FN-coated plates was abolished by treatment of liposomes with endo-F/PNGase-F. Liposomes incorporating alpha 5 beta 1 pretreated with endo-F/PNGase-F also did not bind to FN. When purified alpha 5 beta 1 receptor was treated with endo-F/PNGase-F followed by ZH5 or ZH1, the alpha 5 or beta 1 subunit was precipitated separately, respectively. In contrast, both subunits were always coprecipitated when intact purified alpha 5 beta 1 receptor was directly treated with ZH5 or ZH1. These findings indicate that N-glycosylation of both the alpha and beta subunits of the alpha 5 beta 1 integrin receptor is essential for association of these subunits and for optimal binding to FN.

Identification of novel mutations that confer drug resistance in the human immunodeficiency virus polymerase gene

We generated variants of the human immunodeficiency virus type 1 (HIV-1) that are resistant to 2',3'-dideoxycytidine (ddC) and 2',3'-didehydro-3'-deoxythymidine (d4T) by in vitro selection in MT-4 cells. Portions of flanking protease and integrase sequences as well as the complete reverse transcriptase (RT) open-reading frame of these viruses were cloned and sequenced, using polymerase chain reaction (PCR)-based methods. Mutations were observed at amino acid position 65 (Lys-->Arg; AAA-->AGA) when ddC was employed in the selection procedure and at site 50 (Ile-->Thr; ATT-->ACT) when d4T was used. We confirmed the ability of these mutations to confer diminished sensitivity for these compounds by site-directed mutagenesis, in which these mutations were inserted into the pol gene of infectious recombinant HXB2-D DNA. Viruses that contained the site 65 mutation possessed approximately 5-10 fold resistance against ddC when compared with wild-type HXB2-D. The site 50 mutation conferred approximately 30-fold resistance to d4T in these same assays. Similar results were obtained using primary cord blood lymphocytes in drug resistance assays, indicating that these mutations could confer drug resistance in more than one cell type and that the respective mutations could be expressed in cells of primary origin. No cross-resistance against 3'-azido-3'-deoxythymidine (AZT) was noted for either the site 65 or 50 mutations.

Human cancer cell lines express a negative transcriptional regulator of the interferon regulatory factor family of DNA binding proteins

Members of the interferon regulatory factor (IRF) family of DNA binding transcription factors have roles in growth regulation, antiviral responses, and transcriptional induction of interferon (IFN)-activated early response genes. The IRF family member ISGF3 gamma is the DNA binding component of IFN-stimulated gene factor 3 (ISGF3), a multicomponent complex responsible for the stimulation of IFN-alpha-responsive genes. IFN-alpha-stimulated formation of ISGF3 and subsequent gene expression can be inhibited by phorbol esters or expression of the adenovirus E1A protein. We have investigated IFN signaling in human malignant tumor cell lines of the lung, colon, ovary, cervix, and hematopoietic organs and found some of these cells to be defective for IFN-alpha-induced formation of ISGF3. In many cases, an inhibitory activity termed transcriptional knockout (TKO) correlated with nonresponsiveness. TKO purified from a human papillomavirus-negative cervical carcinoma cell line has a molecular size of 19 kDa. The purified protein interacted with the ISGF3 gamma component of ISGF3, preventing binding of ISGF3 to DNA. Purified TKO displaced ISGF3 from its DNA binding site in vitro and prevented ISGF3 gamma, IRF-1, and IRF-2 from interacting with the IFN-stimulated response element. Partially purified TKO can also directly interact with ISGF3 gamma in the absence of DNA. This protein may be involved with the development of malignancies and the inability of IFN to exert its antiproliferative and antiviral effects.

Identification of a mutation at codon 65 in the IKKK motif of reverse transcriptase that encodes human immunodeficiency virus resistance to 2',3'-dideoxycytidine and 2',3'-dideoxy-3'-thiacytidine

The technique of in vitro selection was used to generate variants of the human immunodeficiency virus type 1 that are resistant to 2',3'-dideoxycytidine (ddC). Most of the pol regions of such viruses, including the complete reverse transcriptase open reading frame and portions of flanking protease and integrase genes, were cloned and sequenced, using PCR-based procedures. Mutations were variously detected at amino acid site 65 (Lys-->Arg; AAA-->AGA) and at a previously reported codon, site 184 (Met-->Val; ATG-->GTG). We introduced the site 65 mutation into the pol gene of infectious, cloned HxB2-D DNA by site-directed mutagenesis in order to confirm by viral replication assay the importance of this site in conferring resistance to ddC. The recombinant virus possessed greater than 10-fold resistance against this compound in comparison with parental HxB2-D. Cross-resistance of approximately 20- and 3-fold, respectively, was detectable against the (-) enantiomer of 2',3'-dideoxy-3'-thiacytidine and 2',3'-dideoxyinosine but not against 3'-azido-3'-deoxythymidine. Combinations of the site 65 and 184 mutations did not yield levels of resistance higher than those attained with the site 65 mutation alone. The presence of the site 65 mutation was confirmed by PCR analysis of peripheral blood mononuclear cells from patients on long-term ddC therapy in 4 of 11 cases tested. Viruses that possessed a ddC resistance phenotype were isolated from subjects whose viruses contained the site 65 mutation in each of four instances. Four of these clinical samples were also demonstrated to possess the Met-184-->Val mutation, and one of them possessed both the Lys-65-->Arg and Met-184-->Val substitutions. Direct cloning and sequencing revealed the site 65 mutation in viruses isolated from these individuals.

Purification and characterization of the U-particle, a cellular constituent whose synthesis is stimulated by Mengovirus infection

We have isolated a cellular protein particle whose synthesis is induced by infection with Mengovirus or TMEV. The U-particle inhibits translation in vitro and binds to both capped and uncapped mRNA's. It is spherical, 12 nm in diameter, and is composed of multiple copies of two polypeptide subunits having molecular weights of 23,000 and 25,000 which do not appear to be glycosylated or phosphorylated. U-particles are capable of inhibiting mRNA translation in vitro.

The cellular U-particle, whose synthesis is induced by mengovirus infection, is homologous to apoferritin

Mengo virus infection of mouse L-cells results in induction of the synthesis of a cellular protein-containing particle, 12 nm in diameter, which was designated U (Boege et al. (1987) Virology 159, 358-367). We have purified the U-particle from virus-infected cells by a series of chromatographic steps and found it to be composed of two polypeptide species (MW 23,000 and 25,000), present in a ratio of approximately 7:3. Neither of these polypeptides is measurably glycosylated or phosphorylated and the U-particle contains no detectable nucleic acid. Several amino acid sequences obtained from CNBr fragments of the U-polypeptides identified them as the H- and L-chains of mouse apoferritin. This finding was supported by immunoblotting and electron microscopy. In terms of function, the U-particle/apoferritin effectively inhibits the translation of mRNAs in reticulocyte lysates. These experiments indicate that apoferritin may perform important functions in eukaryotic cells in addition to iron storage. Finally, we propose mechanisms to explain how Mengo virus infection could specifically induce the synthesis of apoferritin and how increasing amounts of cytoplasmic apoferritin could facilitate virus replication.

Hormone binding and transcription activation by estrogen receptors: analyses using mammalian and yeast systems

We have used affinity labeling, site-directed mutagenesis and regional chemical mutagenesis in order to determine regions of the human estrogen receptor (ER) important in hormone binding, ligand discrimination between estrogens and antiestrogens, and transcriptional activation. Affinity labeling studies with the antiestrogen, tamoxifen aziridine and the estrogen, ketononestrol aziridine have identified cysteine 530 in the ER hormone binding domain as the primary site of labeling. In the absence of a cysteine at 530 (i.e. C530 mutant), C381 becomes the site of estrogen-compatible tamoxifen aziridine labeling. Hence these two residues, although far apart in the primary linear sequence of the ER protein, must be close in the three-dimensional structure of the protein, in the ER ligand binding pocket, so that the ligand can reach either site. Site-directed mutagenesis of selected residues in the ER and region-specific chemical mutagenesis of the ER hormone binding domain with initial phenotypic screening in yeast have enabled the identification of a region near C530 important in discrimination between estrogens and antiestrogens and of other residues important in hormone-dependent transcriptional activation. Some ER mutants with alterations in the carboxy-terminal portion of the hormone binding domain are transcriptionally inactive yet bind hormone and also function as potent dominant negative ERs, suppressing the activity of wild-type ER at low concentrations. These studies reveal a separation of the hormone binding and transcription activation functions of the ER. They are also beginning to provide a more detailed picture of the ER hormone binding domain and amino acids important in ligand binding and discrimination between different categories of agonist and antagonist ligands. Such information will be important in the design of maximally effective antiestrogens. In addition, since there is now substantial evidence for a mixture of wild-type and variant ERs in breast cancers, our studies should provide insight about the bioactivities of these variant receptors and their roles in modulating the activity of wild type ER, and should lead to a better understanding of the possible role of variant receptors in altered response or resistance to antiestrogen and endocrine therapy in breast cancer. In addition, some dominant negative receptors may prove useful in examining ER mechanisms of action and in suppressing the estrogen-dependent growth of breast cancer cells.

Antisense RNA inactivation of gp138 gene expression results in repression of sexual cell fusion in Dictyostelium discoideum

A glycoprotein, gp138, is implicated in the sexual cell fusion of Dictyostelium discoideum. We previously cloned and sequenced the two genes encoding the gp138 protein, GP138A and GP138B (Fang et al. (1993) Dev. Biol. 156, 201–208). Here, we have constructed a vector producing antisense RNA for the gp138 genes and have transformed the vector into Dictyostelium cells. The transformed cells showed a reduction in the amounts of gp138 mRNA and protein and their sexual cell fusion activity was considerably repressed.

[Application of the mathematical model to forecast the incidence rates of seasonal infectious diseases]

The incidence rates of infectious diseases were selected and analysed according to data from the Beijing Railway Area during 1981-1991. We put forward the mathematical model to forecast the incidence rate of dysentery each quarter in 1992. The best mathematical model was selected from analysis of precision, and very useful in the prevention and treatment of seasonal infectious diseases.

Identification of the transcriptional start site of the cyd operon from Escherichia coli

The cydAB operon encodes the two subunits of the cytochrome bd ubiquinol oxidase from Escherichia coli. This enzyme is one of two terminal oxidases in the aerobic respiratory chain of E. coli. It has been demonstrated that expression of the cyd operon is transcriptionally regulated by oxygen via both the fnr and arcA gene products. Whereas arcA clearly serves as a positive regulator at low oxygen tension, there is no consensus concerning the role of the fnr gene product (FNR). In this paper, the transcriptional start site of the cyd operon is identified at position -287 with respect to the translational start. This is located 53 basepairs downstream from the center of dyad symmetry of the putative FNR recognition sequence.

Experimental studies on peripheral nerve regeneration enhanced by nerve growth factor

The effect of exogenous nerve growth factor (NGF) examined on the neural repair of adult rabbit sciatic nerve was evaluated in the present study. A nerve regeneration chamber was created by suturing the proximal and distal ends of a transected sciatic nerve into a silicone chamber. A gap of 6 mm in chamber was left after removal of a 3 mm piece of nerve in the distal ends and insertion of the proximal and distal stumps into the chamber. Animals were operated on bilaterally, one side of the chamber was filled with a 1 mg/ml NGF/normal saline (NS) (experimental) and the contralateral side with NS (control). The regenerated nerves from within the silicone chamber were dissected and fixed 1 to 5 weeks following surgery for histological studies at both the light microscopic and ultrastructural levels. The NGF chamber showed a more mature regenerated nerve based on a larger diameter of the regenerated nerve trunk, a great number of axons, and thicker myelin sheaths.

Molecular cloning and characterization of two genes encoding gp138, a cell surface glycoprotein involved in the sexual cell fusion of Dictyostelium discoideum

The sexual development of cellular slime molds is initiated by acquisition of sexual fusion competence of myxamoeboid cells. During the acquisition of fusion competence in NC4 and HM1 strain cells of Dictyostelium discoideum, opposite in mating types, a glycoprotein gp138, relevant to sexual cell fusion, is expressed on the cell surfaces of both the strains. To investigate the mechanisms controlling cell fusion, gp138 was purified and its N-terminal amino acid sequence was determined. An oligonucleotide with sequence predicted from the amino acid sequence was synthesized to isolate the gene encoding gp138. Two genes, referred to as GP138A and GP138B, were then cloned. They contain the region which encodes the N-terminal amino acid sequence of the gp138 protein and their structure is very similar on the whole. The C-terminal portion of the predictive polypeptides produced by the genes is highly hydrophobic and proline rich. It has homology with a portion of gp80, a glycoprotein for cell-cell adhesion, and PsA, a glycoprotein specific for prespore cells, reported previously in D. discoideum. The expression of GP138A is greater than that of GP138B. The mRNA of GP138A is expressed at the time of acquisition of fusion competence of cells during cultivation but the mRNA of GP138B is not.