Jan 14, 2018 - Center, 1501 Kings Highway, Shreveport, LA 71130. Tel.: 318-674- .... mation of fruiting bodies (Golumbeski and Dimond, 1987;. Coston and ...
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 269, No. 2, Issue of January 14, pp. 146%1476,1994 Printed in U.S.A.
Q 1994 by The American Society for Biochemistry and Molecular Biology, Inc.
Molecular Cloning and Characterization of the Full-length cDNA Encoding the Developmentally RegulatedLysosomal Enzyme @-Glucosidasein Dictyostelium discoideum* (Received for publication, July 1, 1993,and in revised form, September 7,
1993)
John Bush, Jan Richardson$, and James CardelliQ From the Department of Microbiology and Immunology, Louisiana State University Medical Center, Shreveport, Louisiana 71130
The developmentally regulated Dictyostelium discoideum lysosomal enzyme fl-glucosidaseis synthesized as a membrane-associatedglycosylatedprecursor polypeptide which undergoes at least two proteolytic cleavage events to generate a soluble mature lysosomally localized protein. To begin to analyze the mechanisms regulating the sorting of this protein and regulation the during development of the expression of the encoding gene, we have cloned and sequenced a 2.6-kilobase (kb) cDNA which contains a complete 2463-nucleotide open reading frame coding for @-glucosidase. Conceptual translation of this open reading frame predictsa polypeptide similar in molecular mass to the primary translation product of 94 kDa that also contains the same amino acid sequencesof two VS protease derivedpeptides generated from the purified &glucosidase enzyme. The D. discoideum enzyme contained regions highly homologous at theamino acid sequence level to both bacterialandfungal &glucosidases, although these regions did not overlap. A potential cleavable signal sequence was also found in the first 21 amino acids followed by a highly polar stretch of 49 amino acids which (based on amino acid sequencing of the mature @-glucosidase)represents a proregion for this protein. This region is similar in location, size, and charge to the D. discoideum a-mannosidase pro-I region (Schatzle, J., Bush, J., and Cardelli, J. (1992) J. Biol. Chem. 267, 4000-4007). Several small hydrophobic stretches of amino acids were also distributed throughout the protein; however, no obvious transmembrane region(s) were identified which might explain the observed membrane association of the precursor protein. Finally, Northern blot analysis indicated that the gene encoding this enzyme was under developmental regulation. The steady state level of a 2.7-kb&glucosidase mRNA decreased significantly during the aggregation stage of development, from
high levels during growth, and then increased in the form of a larger size 2.8-kb mRNA during the final stages of development.
Mammalian lysosomal enzymes are transported and targeted to lysosomes by at least two different mechanisms (reviewed in Kornfeld and Mellman (1989)). The best characterized lysosomal enzyme targeting mechanism involves the generation of mannose 6-phosphate residues on the Asnlinked oligosaccharide side chains of lysosomal enzymes. These residues serve as recognition markers for mannose 6phosphate receptors which mediate transport of lysosomal enzymes to lysosomes. Mannose 6-phosphate receptor-independent mechanisms for lysosomal/vacuolar targetingin mammalian and yeastsystems have also been described (Kornfeld and Mellman, 1989; Klionsky and Emr, 1990). Dictyostelium discoideum is a haploid eukaryotic organism which appears to use a mannose 6-phosphate receptor-independent mechanism to sort lysosomal enzymes (for review, see Cardelli (1993)). Because it is amenable to both genetic and biochemical approaches, this organism represents auseful system to study alternative lysosomal targeting mechanisms. D. discoideum cells also can be induced by starvation to undergo a relatively simple developmental cycle (Loomis, 1982). Approximately 6 h after development begins, the cells begin to migrate by chemotactically responding to pulses of CAMPand ultimately form aggregates of 10‘ cells. Aggregates then pass through a number of morphological stages ending inculmination that resultsin the formation of amature fruiting body consisting of two types of cells: a stalk of vacuolated cells (stalk cells) which supports a sorus containing dormant spore cells. Three of the better characterized D. discoideum lysosomal enzymes, a-mannosidase, ,&glucosidase,and acid phosphatase are synthesized on membrane bound polysomes, co-transla*This research was supported by National Institutes of Health tionally inserted into the rough endoplasmic reticulum, and Grant DK 39232 (to J. C.) and National Science Foundation Grant glycosylated on Asn residues (Cardelli et al., 1987; Cardelli, DCB-9104576(to J. C.). The costs of publication of this article were 1993). During transport of the membrane-associated precurdefrayed in part by the payment of page charges. This article must sor forms of these enzymes, the mannose-rich carbohydrate therefore be hereby marked “advertisement” in accordance with 18 side chains are sulfated and phosphorylated in the Golgi U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence(s) reported in this paper has been submitted complex (Mierendorf et al., 1985; Cardelli et al., 1986; Freeze, to theGenBankTM/EMBL Data Bankwith accession number(s) 1986). Sulfate and perhaps phosphate are not required for L21014. correct localization of enzymes (Cardelli et al., 1990a; Bush $ Current address: Dept. of Physiology, Dartmouth University and Cardelli, 1990; Freeze et al., 1989). During transport to School of Medicine, Lebanon, NH 03766. lysosomes, these three enzymes are also subjected to several To whom all correspondence should be addressed Dept. of Miproteolytic events which generate the soluble mature form of crobiology and Immunology, Louisiana State University Medical Center, 1501 Kings Highway, Shreveport, LA 71130. Tel.: 318-674- each hydrolase (Mierendorf et al., 1985; Pannell et al., 1985; Cardelli et al., 1986; Bush and Cardelli, 1989). The initial 5756;Fax: 318-674-5764.
1468
~ ~ s o s o~~- aG ll u c o s ~Sequence ~e: and ~ e g u ~ t i o n
FIG. 1. Restriction mapsof
1469
full
length p-glucosidase cDNA and other isolated partial &glucosidase clones. The restriction maps of partial clones and the full-length clone pBG-C and thecorresponding ORF (shoded) are shown. The translational start ATG and s t o TAA ~ codon are also designated. The doked line for pBG-16 represents the length of the non-~-glucosi~se fusion cDNA.
A BG-C H 0.2kb
cleavage event occurs in a Golgi and/or endosomal compartment (Cardelli et al., 1990b;Wood and Kaplan, 1988) and may be required for correct localization of enzymes to lysosomes (Richardson et al., 1988),while final proteolysis occurs within dense acidic lysosomesand is catalyzed by a cysteine proteinase (Cardelli et al., 1989; Richardson et aL, 1988). Although the lysosomal and endosomal vesicles are acidic, low pH is not essential for the correct sorting these hydrolases (Cardelli et aL, 1989). The maintenance of acidic vacuolar compartmen~is, however, required for complete processing of all threelysosomal enzymes.In addition, about 10% of the precursor forms of these enzymes escape cleavage and exit the cell through a constitutive secretory pathway while the mature forms of the enzymes are released from cells in a regulated fashion (Mierendorfet ai., 1985; Pannell et al., 1985; Ebert et al., 1990). The expression of the two lysosomal enzymes,a-mannosidase and @-glucosidase, regulated is during development of D. discoideum (Loomis, 1975). The structural gene encoding amannosidase has been cloned(Schatzle et d., 1992)and shown to be regulated at the level of transcription in response to both a secreted protein termed the presta~ationresponse factor and by starvation (Clarke et aL, 1988, Schatzle et aL, 1991, 1993); however, less is known about the regulation of the expression of the @-glucosidasegene. The rate of &glucosidase synthesis and the steady state level of enzyme activity increases slightly between 1and 3 h into development, decreases significantly by the aggregation stage, and then increases steadily from 18 h of development until final formation of fruiting bodies (Golumbeski and Dimond, 1987; Coston and Loomis,1969). The changes inthe levels of functional @-glucosidase mRNAcorrelates with the rate of enzyme synthesis suggesting that expression maybe controlled at a pretranslational level. TObegin to determine the molecular factors essential for the targeting of @-glucosidaseto lysosomesand theregulation of @-glucosidaseexpression, the cDNA coding for this protein has beencloned and sequenced. The gene is subject to a complex mode of developmental regulation and thepredicted @-glucosidaseprotein has features shared with another D. discoideum lysosomal enzyme,a-mannosidase. MATERIALS ANDMETHODS
and resuspended in cold MES'-PDF buffer (8 mM MES, 0.7 mM CaC12, 0.3mM streptomycin sulfate, pH 6.5). Cells were placed under developmental conditions as previously described (Schatzle et al., 1991). Purification of (3-Glucosidase-Four liters of Ax3 cells at 1-2 X lo7 cells/ml were separated from the medium by centrifugation at 3000 X g for 10 min. The supernatant, which contained 90% of the total &glucosidase enzyme activity, was chromatographed on a Sephacryl 5-300 column (Pharmacia LKB Biotechnology Inc.) as described by Mierendorf et ai. (1983). Fractions containing @-glucosidaseactivity were pooled and applied to an antibody affinity column prepared using monoclonal antibodies specific to f?-glucosidase(Golumbeski and Dimond, 1986) coupled to CNBr activated Sepharose (Pharmacia). The enzyme was eluted with 2 M MgClz and dialyzed against 10 r n NaPO. ~ buffer (pH 6). Fractions containing&glucosidase activity were concentrated by a centrifugation using a Centricell device (30,000M,cut-off, Polyscience) and stored at -20 "C. Purified @-glucosidasewas subjected to SDS-polyacrylamide gel electrophoresis on a 7.5% gel which had been aged 40 h before use (Schatzle et al. 1992). Gel slices containing the 100-kDa mature form of the protein were cut from the gel, inserted into thewells of a 12.5% denaturing polyacrylamide gel, and overlaid with 0.5 pg ofV8 protease in 60 pl of Laemmli (1970) sample buffer. Samples were electrophoresed at 15 mA until the dye front reached the separating gel; electrophoresis was then terminated for 30 min to allow protease digestion of @-glucosidase.Following electrophoresis through the separating gel at 30 m A , peptides were blotted to an Immobilon-P membrane (Millipore Co., Bedford, MA) and stained with Coomassie Blue as previously described (Schatzle et al. 1992). The amino acid sequence was determined from blotted p-glucosidase peptides by liquid phase amino acid sequencing (Core Laboratory Facility, LSUMC-New Orleans, LA). Alternatively, the 100-kDa mature form of the enzyme was subjected to SDS-polyacrylamide gel electrophoresis, blotted to Immobilon-P filters as described above, and subjected to N-terminal sequence analysis. Cloning of Full-length cDNA Encoding a-Glucosidase-AcDNA library was constructed in phage XZAP using poly(A+) RNA from membrane-bound polysomes (Cardelli et al., 1981). Phage were absorbed onto bacterial strains XL-1or BB4 and screened using a panel of nine monoclonal antibodies to &glucosidase (Golumbeski and Dimond, 1986) using the protocol recommended by the m~ufacturer (Promega-Bio~h,Madison, WI). The threeplasmid clones isolated using this protocol were designated pBG-1, pBG-3, and pBG-16. pBG1was sequenced and was found to code for a partial fragment of the &-glucosidasegene (see Fig. 1). The nucleotide sequence of the pBG1and pBG-16 clones was determined using single-stranded DNA and double-stranded DNA following the sequencing protocols of U. S. Biochemical Corp. AI1of the sequencing was done using the dideoxynucleotide chain terminator method and a Sequenase I1 kit (U. S. Biochemical Corp.). To isolate full-length cDNA encoding &glucosidase, the pBG-1 insert was radiolabeled with 32Pand used to screen a different recombinant cDNA library from the one described above constructed in Xgtll with mRNA prepared from cells developing for 4 h (Clontech). Probe hybridization and filter washing were performed as previously described (Schatzle et al., 1992). 29 positive plaques were
Growth and Development of D. discoideum-D. discoideum strains Ax3 (wild-type), Ax4 (wild-type), and Ga2(null mutant) derived from an Ax3 parental strain (a kind gift of Dr. R. Firtel) weregrown axenically in TM medium (Free and Loomis, 1974) at 21 "C in a rotary shaker water bath rotatingat 200 rpm or on SM agar medium at 21 "C in association with K l e ~ s aerogenes. ~ l ~ To initiate devefThe abbreviations used are: MES, 4-morpholineethanesulfonic opment on filters, vegetative amoebae were collected by centri~gation acid; ORF, open reading frame; kb, kilobase.
1470 1 31 61 91 12 1
151 IS1 221 241 271 301 331 361 CGTTCCAGMTUCGT-T V P E S R L
391
D
421 451 451 511 541 571 651 631
- ~ ~ T C T T T ~ A T A ~ ~ M ~ A T A C C M ~ C G C ~ T A G ~ ~ T ~ C C M ~ ~ T ~ T 691
F
G
D
G
L
S
Y
T
T
F
N
.
1
Y
T
N
L
A
C
S
N
C
X
P
1
S
G
Q
S
G
N
1
Y
781
811 TMAGATT
FIG. 2. The completeDNA sequence of the @-glucosidasecDNA pBG-C; GenBank" accession no. L21014.The numbers to the left designate the amino acid positions from 1 to 821 in the ORF. The seven po~entialN-linked giycosylation sites aredenoted by dark closed circles while two large predicted hydrophobic stretches are contained withinboxes. The two @-glu~sidase peptide sequences determined from sequencing purified glucosidase are u ~ e r ~ i The ~ dpredicted . signal sequence cleavage site (amino acid position +24) i s represented by a dark triangte and the predicted pro region (+24 to f72) is ~ ~ eby dashed r l lines. ~ The ~ circkd ~ amino acids represent consented residues which probably make up the active site of the enzyme. The arrowheads point to the N-terminal amino acids found in purified mature @glucosidase. isolated, and partial DNA sequencing and restriction mapping in&cated that 20 of these clones were related to each other. One clone designated as pBG-C was further characterized and found to contain three EcoRI fragments 1.6, 0.8, and 0.2 kb in size which were then subcloned and subjected to single-stranded dideoxy sequencing. Protein Analysis-The predicted @-glucosidaseamino acid sequence was analyzed using several computer programs ( ~ a c v e c t o r
version 3.5,Macintosh model SE30computers) designed to determine protein secondary structure, surface probability, and hydrophobicity. Proteins homology searches were conducted using a computerized version of the Lipman-Pearson and Wilbur-Lipman high speed protein library searches (NationalBiomedical Research Foundation data base, release number 31.0). S o u € h e~~ ~ ~ A ~ ~ s ~ - G eDNA n o was m i isolated, c digested with
1471
I
I
I
I
1 1
I
1
I
I
I
I
1
1 I
I1 '
I
1 I
1'
I
I
I
I
I
I
I
I
I
I
I
I
100
200
300
400
500
600
700
800
I
I
I
I
I
1
300
400
500
I
1
I
I
100
200
I
600
700
I
800
FIG. 3. Organization of the predicted &glucosidase precursor. The three predicted domains of the @-glucosidaseprotein are illustrated in the figure and are designated as the signal sequence, propeptide, and mature protein. Additionally, two large hydrophobic regions (checked boxes) and the predicted active site (dotted box) of the protein are also noted. The amino acid positions from 1 to 821 are also shown below the topschematic of the protein. Potential glycosylation sites are indicated by the lollipop symbol. The hydrophilicity profile of 8-glucosidase was calculated by the Kyte-Doolittle method (1982) based on a window setting of 7. The secondary structure profile of the protein is positioned at the bottom of the diagram. The combined method of secondary structure calculation uses both the Chou-Fasman (1978) and Robson-Garnier (Garnier et ai., 1978) computer programs. TABLE I Comparison of ~ g from ~different ~ organisms ~ ~ ~ e ~ References for sources of @-glucosidaseare asfollows: LJ.discoideum (this publication); C. t ~ r m o c e ~(Grabnitz ~m et at., 1989);I(. marxianus (Raynal et al., 1987);H.anomla (Kohchi and Toh-e, 1985); R. albus (Ohmiya et al., 1990), and S. commune (Moranelli et al., 1986). Organism
Molecular mass
PI"
identity/ Percent similarityb
Region of homology'
kDa
D.discoideum 4.77 Bacteria, C. thermocellum Yeast, K.marxianus 5.17 Yeast, H.anomla4.70
89.4 84.1 93.9 89.5 104.3 NDd
5.44
Bacteria, Ruminococcus 4.84 albus Fungus, S. commune ND a Isoelectric point. Percent identitylsimila~tywas calculated as described under "Materials and Methods." e Region of homology refers to D.d ~ c o ~ @-glucosidase e ~ m amino acid positions, Not determined. a variety of restriction enzymes, and subjected toSouthern blot analysis as described by Schatzle et al. (1992). Blots (Genescreen Plus, DuPont NEN) were hybridized in 1XP at65 "C and washed as described above for the filter lift procedure. BG-1 was nick-translated (Life Technologies, Inc.) and used as a probe in both Southern and Northern blot analysis as previously described (Schatzle et al., 1992). RNA Extractions and Northern Blots-For each developmental or
29/46 29/42 21/36 24/41 23f35
453-821 138-363 168-736 517-695 186-486
tometer (LKB 2202 ultrascan, LKB Inc., Gaithersburg, MD). To normalize the levels of @-glucosidasemRNA to ribosomal RNA, the blots were stained with methylene blue and quantified as previously described (Schatzle et al., 1991). RESULTS AND DISCUSSION
1991). Filters wire exposed to Kodak%kR-5 film a t -70 'C, and @: different monoclonal antibodies, all capable of recognizing glucosidase mRNA levels were quantified by a scanning laser densi- mature &-glucosidase onWestern blots (Golumbeski and Di-
Lysosomal 8-Glucosidase:Sequence and Regulation
1472
Development In Suspension
Development Fllterron 2.8 kb 2.7 kb
D.r.,opm.nl
0
1 2
1 4
6 I 14 1 6 1 0 2 0 2 2 24
0
1 2 3 4
I I 14 1818 20 22
0
1a
22
24
0
0 2 mRNA
FIG. 4. 8-Glucosidaee mRNA levels are regulated differently in cells developing on filters uersw Ax3 cells developing in suspension. Total RNA was extracted at theindicated time points and subjected to Northern blot analysisusing radiolabeled @-glucosidase or D2 DNA inserts as probes (see “Materials and Methods”). Thetop three panelsare the autoradiographs using the@-glucosidaseinsert as a probe during a developmental time course where cells were in suspension (first panel)or plated on filters (second panel).The third panel (extreme right)shows the two mRNA@-glucosidase speciesof 2.7 and 2.8 kb found a t T = 0 and 22/24 h of development. The bottom panels reurobed with the D2 DNA probe revealing the presence of the 1.8-kb D2 show the autoradioPraDhs of the same two blots shown on the left mRNA during thetwo’time courses.
Nucleotide Sequence and Deduced Amino Acid Sequence of mond,1986). The expression library was generatedusing mRNA isolated from membrane-bound polysomes previously the @-GlucosidasecDNA, BG-C-The 2594-nucleotide pBG-C shown tobe enriched for mRNAs encodinglysosomal enzymes cDNA insert wassequenced on both strands as described (Mierendorf et al., 1983). Threepositive clones were purified under “Materials and Methods.” A single long ORF of 2463 and found to contain inserts (Fig. 1) of 1.1 kb (pBG-l), 1.3 nucleotides coding for821 aminoacids wasdetected extending kb(pBG-3),and 1.8 kb (pBG-16). Southernblotanalysis from nucleotide 68 to nucleotide 2594. The ORF begins with indicated that the clones pBG-1 and pBG-16were related an ATG (nucleotides 67-69) and ends witha TAA stop codon suggesting it coded for a full-length protein. This first ATG (they cross-hybridized understringentconditions),and Northern blot analysis indicated that both of these cDNAs most likely represents the initiation codon because it is prerecognized asingle mRNA species of2.8 kb(resultsnot ceded by a veryA/T-rich region containing stopcodons in all codons are usually shown); the BG-3clone which hybridized to a 2.0-kb mRNA three reading frames; translation initiation was not studied further. 2.8 kb is the appropriate size expected preceded by a string of As in D. discoideum. Conceptual translationof this long ORF yielded a predicted for a mRNA encoding a protein of 94 kDa, the molecular mass of unglycosylated @-glucosidase. Southern blot analysis protein of 89-kDa molecular mass and an isoelectric point indicated that an antisense oligonucleotide synthesized based (PI) of4.8. This is close to the molecularmass of the @on the aminoacid sequence of a V8 protease-released peptide glucosidase primary translation product of 94 kDa (Cardelli et al., 1987) and the experimentally determined PI of the from pure @-glucosidase hybridized to both cDNAs (results predicted amino acid senot shown).Finally, nucleotide sequenceanalysis of these two protein (results not shown). The clones indicated they were identical where they overlapped, quence also matched the sequence of the two V8-derived @and each contained an ORF coding for an incomplete poly- glucosidase peptides (represented by the single underlines in peptide homologous to @-glucosidasefrom the yeastKluyuer- Fig. 2) indicating the pBG-C clonecodesfor authentic @omyces marrianus (Raynal et al., 1987). This suggested that glucosidase. Finally, seven possible N-glycosylation sites were these clones represented incomplete cDNAs encoding D. dis- located throughout the lengthof the polypeptide (Figs. 2 and coideum @-glucosidase. Unfortunately, pBG-16 (containing 3) whichagrees well withtheexperimentallydetermined number of six N-linked oligosaccharide side chains per prethelargest cDNA insert)represented a fusioncloneand contained the extra DNA sequencefromnucleotide 1-216 cursor (Cardelli et al., 1986). The predicted @-glucosidaseamino acid sequence was used (represented by the dashed line in Fig. 1) which coded for a stretch of amino acids found to be 50% identical to yeast to search the NBRF protein sequence data base to identify homologous proteins. The search revealed that the polypepsteroyl-CoA desaturase (results not shown). To isolatefull-length cDNAsencoding @-glucosidase, a tide contained long stretches of amino acidsequence that second Xgtll library prepared using mRNA fromcells devel- were significantly homologous to @-glucosidase amino acid oping for 4 h was screened using the pBG-1cDNA insert as sequences identified from both a yeast species, K . marxianus a radiolabeled probe. Twenty-nine clones containing inserts (Raynal et al., 1987), and a bacterial species, Clostridium ranging insize from 0.2 to 2.6 kb were purifiedto homogeneity, thermocellum (Grabnitz et al., 1989) (Table I). Interestingly, and partialDNA sequencing and restriction enzyme mappingthese homologous sequences did not overlap in position beindicated that20 of these clonesoverlapped with pBG-1. The tween the yeast and bacterial proteins. For instance, the D. @-glucosidase(29% cDNA insert from thelargest clone, pBG-C (Fig. l), was discoideum region homologous to the yeast amino acid sequence identity and46% similarity) was located sequenced on both strands, and, as described in the next in the N-terminal portion of the D. discoideum protein and section, thiscDNA coded for full-length @-glucosidase. region Southern blot analysis of genomic DNA digested with a spanned aminoacid positions 138-363. In contrast, the variety of restriction enzymes indicated that pBG-C recog- homologous to Clostridium @-glucosidase (29%amino acid the C-terminal nized only single DNA fragments when used as a probe under identity and 46% similarity) wasfoundin portion of the slime mold protein (amino acid positions 453moderatelystringent hybridization conditions(resultsnot shown). This is consistent with previous genetic data which 821). The D. discoideum @-glucosidaseamino acid sequence indicate that D. discoideum @-glucosidase is encoded by a was also homologous to a lesser extent to@-glucosidasesfrom other species (TableI). However, no significantsequence single gene named gluA (Loomis, 1980).
Lysosomal P-Glucosidase: Sequence and Regulation
1473
homology wasshared between the D. discoideum @-glucosidase 2 and 5 followed bya stretchof 19 uncharged amino acids (11 and human glucocerebrosidase (results not shown). Interest- which were hydrophobic). Based on the rules proposed by von ingly, the predicted molecular masses and isoelectric points Heinje (1984) we predict that cleavage of the signal sequence of the @-glucosidaseenzymes from the six listed organisms occurs on the carboxyl side ofGly24. A very hydrophylic (Table I) were very similar (ranging from 89 to 104 kDa in stretch of 47 amino acids extending from amino acid positions molecular mass and 4.7 to 5.4 in isoelectric point, respec- 25-71 immediately followed the putative signal sequence (dotted underline in Fig. 2; striped box in Fig. 3); 15 charged amino tively). Predicted Structural Featuresof the @-Glucosidase Polypep- acids and 14 polar amino acids were distributed throughout tide-Fig. 3 indicates a schematic diagram of the predicted this span. The 105-kDa @-glucosidaseprecursor is cleaved domains of the D. discoideum @-glucosidasepolypeptide along intracellularly first in a Golgi/endosomal compartment to a with a computer assisted analysis indicating hydrophobic 103-kDa intermediate form and then in lysosomes to a 100regions and regions predicted to have secondary structure. kDa mature form. Since the molecular mass of this hydroThe most hydrophobic region was found at the N terminus phylic stretch of amino acids is approximately 5 kDa, we (amino acid positions 4-21) and contained features indicating reasoned that itmight be a “pro”region subject to proteolytic that itmay function as asignal sequence (Fig. 2). For instance, removal. To directly testthis hypothesis, we purified the positively charged lysines were found at amino acid positions mature 100-kDa @-glucosidaseproteinas described under “Materialsand Methods” and determined the N-terminal amino acid sequence for 8 residues.Two Ntermini were A 1 .o present in the purified preparation and corresponded to Ile7’ and Ser74.Since the processing of the 103-kDa intermediate form of @-glucosidaseis catalyzed by a cysteine proteinase(s), 0-0 Suspension 0.8 we propose that theactual cleavage site generating the mature subunitis at the L y ~ ~ ~ L ypair s ‘ jand ~ that an N-terminal peptidase generates the two different mature N termini.How0.6 ever, we cannot rule out the possibility that these N termini were generated artifactually duringpurification of the protein. The processing of the 105-kDa precursor to the 103-kDa 0.4 intermediate form is catalyzed by a proteinase having properties similar to yeast kex2 and theLys4’Ar$’ pair represents a possible cleavage site for this processing event (Richardson 0.2 et al., 1988). Interestingly, The D. discoideum a-mannosidase lysosomal enzyme also contains a short N-terminal hydrophylic pro region contiguous with the signal sequence that 0.0 has been demonstrated by direct amino acid sequence of the 0 2 4 6 8 1012141618202224 mature 60-kDa a-mannosidase subunitto be removed in vivo. Time of Development (Hours) This pro region also contains an amino acid pair that should B be recognized bya kex2 proteinase (Fuller et al., 1989). Interestingly, the size, location, and charge properties of D. discoideum a-mannosidase and @-glucosidasepro regions are simiC H J Suspension lar to both yeast and plant hydrolase pro regions that have 0.8 been found to be necessary and sufficient for localization to e 0 lysosomes (Klionsky and Emr, 1990; Matsuoka and Nakamura, 1991). Although it is reasonable to propose that these pro regions play a role in targetingof hydrolases to lysosomes in Dictyostelium, it should be noted that the targeting information for the slime mold @-hexosaminidaseenzyme does not reside in a pro region (LaCoste et al., 1992; Graham et al., 1988) (in fact @-hexosaminidaseis not cleaved at all during transit tolysosomes). 0. In addition to the putative signal sequence, the @-glucosidase polypeptide contained two other hydrophobic stretches greater than 14 amino acids in length (boxed amino acids in 0. Fig. 2, and hatched boxes in Fig. 3). The first region may form 0 2 4 6 8 1012141618202224 a @-sheetstructure thatextends from amino acid position 216 Time of Development (Hours) to amino acid position 235 (with a hydrophobic index greater than 2.0). The second region extends from amino acid position MORPHOLOGY A 385-416 and also contains subregions with hydrophobic inSTAGE Preagg---Aggregation---Slug--Culmination--Completion dices greater than 2.0. Unlike the first region, the region may FIG. 5. Graphical representation of laser densitometric assume an a-helicalstructure.These two regions maybe scans of the autoradiographs in Fig. 4. Panels A and B illustrate responsible in part for the association observed in vivo bethe relative levels of P-glucosidase (in suspension, open circles, or on tween the precursor of @-glucosidaseand intracellular memfilters, closed circles) and D2 mRNA (in suspension, open squares, or on filters, closed squares), respectively, as determined by laser densi- branes; a-mannosidase also contains hydrophobic stretches tometric analysis of autoradiographs and normalized to the levels of that may function to anchor the precursor to membranes. In ribosomal RNA in each lane of the probed Northern blots (Schatzle theory, both of these regions are long enough to span the lipid et al., 1993). bilayer, however, neithera-mannosidase nor @-glucosidase
*-9
Lysosomal /3-Glucosidase: Sequence and Regulation
1474
Hours
Of
Development CsA
Ga2
NC4 Ax3
Ax4
0 2 4 10
2 4 7 1 5
0
- ”+
”++
” “ +
2
6 12
0
2
6 1 2
””
FIG. 6. &Glucosidase mRNA levels decrease in cellsthat become aggregation competent. Total RNA was extracted at the time points in development indicated below each panel from the D. discoideurn strains, NC4, Ax4, Ax3, and Gn2 and subjected to Northern blot analysis using radiolabeled 8-glucosidase DNA as a probe. The aggregation competence state of the cells is shown below each time point as either CsA (aggregating) or CsA - (non-aggregating).
+
Cells Grown:
I l l Axenically
, ,, ,
(F
Hours of Development
0
2 4
7 1 5 2 22 6
j
Bacterial Food Source
0
I
2 4 7 1517 2 2 2 6
FIG. 7. The pattern of &glucosidase expression is the same during development regardless of how cells weregrown prior to the initiation of development. Ax3 cells which were grown axenically or on bacterial-seeded plates were harvested and platedon filters to initiate development. Total RNA was extracted at the indicated time points and subjected to Northern blot analysis using radiolabeled 8glucosidase DNA as a probe. The two panels are the autoradiographs from one set of these experiments. The left panel shows the levels of pglucosidase mRNA in cells previously grown axenically while the right panel indicates the mRNAlevels in cells grown on bacteria.
and are likely to be integral membrane proteins (Cardelli, 1993) beginning aggregation).Cellscompletedaggregation suggesting these hydrophobic regions probably do not cross formed tight mounds by T = 12 h. Steady state levels of @glucosidase mRNA remained very low until the culmination lipid bilayers. The active siteof @-glucosidasefrom Aspergillus wentii has stage (7‘ = 16-18 h) of development a t which time a greater been identified (Bause andLegler, 1980) and shown to containthan 50-fold increase was observed Fig. 5 is a laser densitoan invariant Asp separated from a conserved Glu 10 amino metric scan of the autoradiographs of Fig. 4 to quantify the well with published reports (Golacids upstream. Fig. 2 indicates the amino acid residues which data. These data correspond might form the active siteD.indiscoideum @-glucosidase.T r p umbeski and Dimond, 1987; Loomis, 1975) on the regulation and Gly amino acid residues also appear to be conserved in of levels of @-glucosidase enzyme activity, although the inaddition to theAsp and Glu residues for @-glucosidase froma crease in mRNA during earlydevelopment reported here was variety of organisms including D. discoideum (Fig. 2), Hun- not as dramatic as previously reported. Interestingly, the @glucosidase mRNA that reappeared late in development was senulaanomala(KohchiandToh-e, 1985),Schizophyllum 100-150 nucleotides largerthanthe @-glucosidase mRNA commune (Moranelli et al., 1986), K. marxianus (Raynal et present early in development (see far right panel of Fig. 4 al., 1987) and C. thermocellum (Grabnitz et al., 1989). which indicates mRNA from three late time points bracketed Developmental Regulation of the @-GlucosidaseGene-Previous studies indicated that therelative levels of translatable by RNA from vegetative cells).The reason for the differences @-glucosidasemRNA and enzyme activity are high in growing in size of the mRNAs is presently unknown butmay include cells and increase during early development, followed by a alternative post-transcriptional processing and/or use of aldecrease to negligible levels during aggregation and slug mi- ternative promoters. In contrast to the results described above gration, and a return to high levels during the final stagesof for cells developing on filters, steady state levels of @-glucodevelopment (Costonand Loomis, 1969; Golumbeski and sidase mRNA dropped 50% in the 1st h and then decreased 20 hin Ax3 cells suspended ina Dimond, 1987). T o more accurately determine the changes in slowly duringthenext @-glucosidase geneexpression during development, RNAwas buffered salt solution(Figs. 4 and 5 ) ; only a slight increase in size of the mRNA wasobserved extracted from growing cells and from cells a t various times the level and no change in the during development, and subjected to Northern blot analysis. after 20-24 h of starvation. The Northern blots described above were stripped of DNA The “P-radiolabeled BG-1 insert was used as a probe. Our laboratory Ax3 wild-type cells were allowed to develop under and reprobed with a second radiolabeled cDNA, D2, whose two different conditions: 1) on filters underwhich they com- encoding gene is transcriptionally inducedin cells actively plete the entiredevelopmental cycle and form fruitingbodies sending and receiving CAMP pulses, a developmentally inand 2) in a buffered salt solution where normallycells reach duced signaling pathway required for chemotaxis and aggreonly the aggregation competent stage. Fig. 5 indicates that gation. Levels of D2 mRNA rapidly increased at thebeginning high levels of @-glucosidasemRNA were observed in growing of aggregation ( T = 4-6 h) in cells developingon filters (Figs. cells and in cells during early development on filters( T = 0- 4 and 5). In contrast, our Ax3 wild-type cells suspended in a salt solution contained significantly lower levels of accumu4 h). The levels of mRNAdramatically decreased at the by 6-8 h, suggesting these cells did not beginning of the aggregation stage ( T = 6-8 h) of development latedD2mRNA (the lawn of cells had begun to “ripple,” an indication of completely become aggregation competent.This was con-
~ y s o s o ~ u~Z- G l u c o s ~Sequence ~e: and ~ e g u ~ ~ ~ ~ n
firmed by an experiment which indicated that Ax3 cells in suspension (in contrast to cells on filters) accumulated only low levels of the cell surface glycoprotein gp80 (also termed contact sites A or CsA) which is required for cell adhesion during aggregation and accumulates at thattime (results not shown). Also, cells recovered from suspension did not form polar end to end contacts (typical of aggregation competent cells) nor did they form aggregation streams when plated on plastic dishes (results not shown). These data suggest that the decline in steady state levels of @-glucosidasemRNA depended on cells becomingag~gation-competent,and that our laboratory Ax3 strain did not reach this stage of development when starved in a buffered salt solution. To further test this, two other wild-type strains, Ax4 and NC4, shown by others to be capable of becoming aggregation competent in suspension (Loomis, 1982), were placed under the developmental conditions described above.Northern blot analysis (Fig. 6) of RNA isolated from these strains during development in suspension indicated that at the time ( T = 8-10 h) cellsbecameaggregation competent (Le. CsA positive), 0glucosidase mRNA levels rapidly dropped. Furthermore, 8glucosidase mRNA levelsdid not decrease in a mutant strain deleted of the Ga2 gene (Kumagai et ai., 1991) when cells were starved on filters (Fig. 6 ) or in suspension (results not shown). Ga2 is a subunit of a heterotrimeric G protein complex that is functionally coupled to the cyclic AMP receptor (cAR1) and is necessary for cAMP signal transduction required for aggregation.Therefore, the decrease in levels of pglucosidase mRNAdepended on cells acquiring the ability to respond to cAMP signals and becoming aggregation competent regardless of whether this event occurred in suspension or on filters (i.e. development on filters is not an absolute requirement to affect a decrease in the level of @-glucosidase mRNA). Moreover, the increase in concentration of a larger size 8-glucosidase mRNA occurring at approximately 16-18 h of development may depend on the ability of cells to enter the culmination stage which only occurred in cells developing on filters. The developmentally induced decrease in thelevels of certain D. discoideum mRNAs depends on the prior growth conditions of the cells. For instance, in cells grown axenically, the steady state levels of mRNAs coding for ribosomal proteins remain constant until developing cells reachthe aggregation stage at which time mRNA levels rapidly and coordinately decrease (Steel and Jacobson, 1987). In contrast, in cellsgrownin association with bacterial as a food source, levels of ribosomal protein mRNAs rapidly decrease upon the initiation of development (Singleton et al.,1989).As indicated in the Northern blot ofFig. 7, @-glucosidase mRNA levels decreased at 4-7 h of development (beginning of aggregation) and increased again at 22 h of development (culmination stage) regardless of whether cells were grown axenicallyor in association with bacteria. Additionally, regardless of the conditions used to grow cells both the 2.7- and 2.8-kb @-glucosidase mRNAs were observed at the appropriate times during development. This suggests that prior growth conditions do not influence either the timing of the decrease and subsequent increase in levels of both forms of the @-glucosidase mRNA and that themechanisms regulating the decrease in levels of @-glucosidase mRNA are different than the mechanisms regulating expression of the ribosomal protein. CONCLUSION
We report here the cloning and DNA sequence analysis of the cDNA encoding the full-len~h precursor polypeptide to
1475
lysosomal @-glucosidase from D. discoideum. The predicted protein contained regions homologousto both fungi and bacterial @-glucosidasesand also contained a potential N-terminal signal sequence followed bya short very hydrophylic “pro” region similar in location and size to the pro-I domain of D. discoideum a-mannosidase. Short N-terminal localizedhydrophylic pro regions function to target hydrolases to lysosomes in plants and yeast cells. The @-glucosidaseprecursor also contained hydrophobic regions that might facilitate association of the protein with intracellular membranes. The expression of the gene encoding lysosomal @-glucosidase was under a relatively complex mode of regulation during develop men^ the relatively high steady state levels of mRNA in growing and early developing cells decrease by the aggregation stage of development to undetectable levels. This reduction in mRNA levels depended on cells acquiring the ability to aggregate. This was followed by a significant increase during the final stages of development in the levels of a @-glucosidase mRNA larger in size than thatobserved in early development. Acknowledgments-We thank Dr. Richard Firtel for the D2 cDNA and theGa2 mutant cell lines. We also thank membersof the Cardelli laboratory for critical reading ofthe manuscript, and we thank Karl FranekandDr.Mike Wolcott for aiding in the purification of 8glucosidase. We recognize the support of the LSUMC Centers for Excellence in Cancer Research, and for Excellence in Arthritis and Rheumatology. REFERENCES Bause, E., and Legler, G . (1980) Biochim. Biophys. Acta 626,459-465 Bush, J., and Cardelli, J. A. (1989) J. Biol. Chern. 264,7630-7636 Bush, J., Ebert, D. L., and Cardelli, J. (1990) Arch. Biochem. Biophys. 283, 158-166 Cardelli, J. A. (1993) in Endosomes and Lysosomes: A Dynnmic Relationship (Stoee, B., and Murphy, R.! eds) pp. 341-390, JAI Press, Greenwich, CT Cardelh, J., Knecht, D., andDlmond, R. (1981) Deu. Bwl. 82,180-185 Cardelli, J. A., Knecht, D. A,, Wunderlich, R., and Dimond, R. L. (1985) Deu. BioL 110,147-156 Cardelli, J., Golumbeski, G., and Dimond, R. (1986) J. Cell Bwl. 102, 12641270 Cardelli J. A., Golumbeski, G. S., Woychik, N. A., Ebert,D. L., Mierendorf, R. C., and Dlmond, R. L. (1987) Methods CeU Bioi 28,139-155 Cardelli, J., Richardson, J., and Miears, D. (1989) J. Bwl. Chem. 2 6 4 , 34543463 Cardelli, J. A., Bush, J., Ebert, D., and Freeze, H.H. (199Oa) J. Bkl. Chem. 266. ..,W7-R8F13 - - .. .- - Cardelli, J., Schatzle, J., Bush, J., Richardson, J., Ebert, D., and Freeze, H. (l99Oh) Deu. Genet. 11,454-462 Chou, P., and Faaman, G. (1978) Annu. Rev. Biochem. 47,251-276 Clarke, M., Kayman, S., and Yang, J. (1988) Deu. Genet. 9,315-326 Coston, M., and Loomis, W. (1969) J. Bacterial. 100,1208-1217 Ebert, D., Jordan, K., and Dimond, R. (1990) J. Cell Sci. 96,491-500 Free, S., and Loomis, W. (1974) Bioehimie 66,1525-1528 Freeze, HI. (1986) Mol. Cell. Bioehem. 72,47-65 Freeze, H.,Koza-Taylor, P., Saunders, A., and Cardelli, J. (1989) J.Bwl. Chem. 264,19278-19286 Fuller, R.,Brake, A., and Thorner, J. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 1434-1438 Gamier, J., Osguthrope, D., andRobson, B. (1978) J. Mol. Biol. 120,97-120 Golumbeski, G., and Dimond, R. (1986) A d . Biochem. 164,373-381 Golumbeski, G. S. and Dimond, R. L. (1987) Deu. Biol. 123,494-499 Grabnitz, F., Rucknagel, K., Seib, M., and Staudenbauer, W. (1989) MOL & Gen. Genet. 217,70-76 Graham, T., Zassenhaus, H., and Kaplan, A. (1988) J. Bioi Chem. 263,168232”
KIiiii. Bio/. dhem. 268,19632-i9&39 Sin leton, C., Manning, S., and Ken, R. (1989) Nucleic Acids Res. 17, 96799f92 Steel, L., and Jacobson A. (1987) Mol. CeN Bid. 7,965-972 van Heijne, G. (1984) d Mol. Bid. 173,243-251
