Feb 5, 2019 - 20892 and the VDepartment of Biochemistry and Molecular Biology, I( Division of ... Should a yeast protein that is 70% identical to a human.
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Vol. 266, No. 4, Issue of February 5, pp. 2606-2614, 1991 Printed in U.S.A.
Human ADP-Ribosylation Factors A FUNCTIONALLY CONSERVED FAMILY OFGTP-BINDINGPROTEINS* (Received for publication, July 9, 1990)
Richard A. Kahn$#, Francis G. Kernll, Jenny Clark$, Edward P. GelmannII, and Cherrie Rulka$ From the $Laboratory of Biological Chemistry, Division of Cancer Treatment, National Cancer Institute, Bethesda, Maryland 20892 and the VDepartment of Biochemistry and Molecular Biology, I(Division of Medical Oncology, Vincent T. Lombardi Cancer Research Center, Georgetown University Medical Center, Washington, D. C. 20007
A newmember,hARF4, of the ADP-ribosylation binding proteins which will ultimately be found expressed in factor (ARF) family, a subset of the superfamily of a single cell type or tissue will likely total between 30 and regulatoryGTP-bindingproteins,hasbeencloned 100. However, as theonly known activities of the vast majority from a cDNA expression library. Two other human of these proteins are guanine nucleotide binding and hydrolARF cDNA sequences, designated human ARFl and ysis, it is currently not possible in many instances todiscrimare 96% inate species differences from subfamily differences. Thus, ARF3, have been reported previously and identicalinaminoacidsequence.Ahuman ARFl attempts at grouping theseproteinsinto subfamilies or cDNA, significantly longer than previously described branches of a phylogenetic tree hasrelied solely onstructural clones, was obtained, by cross-species hybridization information obtained from cDNA-derived protein sequences. using a bovine ARFl cDNA probe. Bovine ARFlp and human ARFlp are 100%identical while each is only Should a yeast protein that is 70% identical to a human 80% identical to hARF4p. Thus, hARF4p is the most protein be considered a homologue or a member of a distinct divergent of the mammalian ARF proteins identified. family? If there are two proteins in the same species that are Northern blot analysis revealed the expression of at 80% identical do they represent one or two subfamilies of least three differentARF messages in human placenta proteins? Such questions can ultimately only be answered by functional data generated from both in vitro biochemical data and adrenal carcinoma cells. Both hARFl and hARF4 encode GTP-binding pro- and in vivo analysis of function. The ADP-ribosylation factor (ARF)’ was the first of the teinswithpredictedmolecular masses of 20,0002 1,000Da. Biochemical analysis of the purifiedrecom- superfamily of smaller, monomeric GTP-binding proteins pubinant proteins revealeda high degreeof conservation rified from tissues (1, 2) and remains one of the few such of nucleotide binding properties and in vitro ARF ac- proteins with a defined i n vitro activity, other than nucleotide tivities. binding or hydrolysis. ARF serves as the protein cofactor ARF is an essential genein theyeast, Saccharomyces required for efficient ADP-ribosylation of the stimulatory recerevisiae, and is encoded bytwo genes. Expression of gulatory subunit of adenylate cyclase, Gs, by cholera toxin (3, either hARFlp or hARF4p in yeast was found toforrescue recent review see Ref. 4). This activity is dependent on the the lethaldouble mutant, arfl-arf2-, thus demonstrat- binding of activating guanine nucleotides (GTP, Gpp(NH)p, ing the functional conservation of ARF functions be- GTP+) to ARF ( 5 ) . The binding of activating guanine nutween yeast and man. The combination of in vivo and cleotides has also been shown to induce a conformational in vitro assays for ARF function provides a specific change in ARF, as evidenced by coincident changes in intrinand unambiguous meansof determining bona fide ARF sic fluorescence and GTP binding ( 5 ) . Thus, GTP is an proteins from divergent species from among the rap- allosteric activator of ARF, p21 ras, and the trimeric GTPidly increasing number of structurally related, small binding regulatory proteins. molecular weight GTP-binding proteins. ARF activity and immunoreactivity have been detected in every eukaryotic cell or tissue examined, including those from man, mouse, turkey, fruit flies, slime mold, yeast, and the Technical advances in the areas of molecular biology and plant, Arabidopsk (6, 7). Cloning of ARF cDNAs by hybridiprotein fractionation have allowed the identification of a zation with nucleotide probes has identified at least two ARF surprisingly large number of structurally related monomeric genes in cows (designated bARFl (8) and bARF2 (9)), man GTP-binding proteinswith molecular masses between 20 and (designated hARFl (10) and hARF3 ( l l ) ) , and yeast (desig25 kDa. The identification of many of these proteinshas come nated yARFl (7) and yARF2 (8)).In each of these cases the in large part from low stringency hybridizations of cDNA two genes from each organism were 96% identical, while yeast libraries with probes derived from previously cloned proteins, and mammalian ARFs are each only about 74% identical. primarily ras p21. The total number of the smaller GTP- Each gene encodes a GTP-binding protein of 180 or 181amino acids. * This work was supported in part by Public Health Service grant ARF is an essential protein in the yeast (Saccharomyces CA-50376from the National Cancer Institute (toF. G . K.). The costs cerevisiae and is encoded by two genes, A R F l and ARF2 (7).
of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. The nucleotide sequence(s)reported in this paper has been submitted to the GenBankTM/EMBL Data Bank withaccessionnumber(s) M36340 and M36341. 4 To whom correspondence should be addressed.
’
The abbreviations used are: ARF, ADP-ribosylation factor; mARF, mammalian A R F bARF, bovine ARF; hARF, human ARF; yARF, yeast (S. cerevisiae) ARF; GTP+, guanosine 5”thiotriphosphate; SDS, sodium dodecyl sulfate; bFGF, basic fibroblast growth factor; hp, base pair(s); kb, kilobase pair(s); HEPES, 442-hydroxyethyl)-1-piperazineethanesulfonicacid.
2606
ADP-Ribosylation Human Disruption of ARFl results in three phenotypes, slow growth, cold sensitivity, and increased sensitivity to fluoride. Loss of ARF2 has no discernible phenotype, while the double ARF disruption is lethal (7). It is likely that thegenetic differences between mutations in ARFl and ARF2 are due to differences in the level of expression of the two proteins, as yARFlp accounts for about 90% of total ARF protein in the cell. Overexpression of yeast ARF proteins hasalso been shown to slow or arrestgrowth of yeast cells (7). A role for ARF proteins in the protein secretory pathway has recently been described, largely from the phenotype of yeast ARF mutants and genetic interactions with previously characterized SEC genes (12). Consistent with this role in secretion, ARF is highly concentrated in the Golgi apparatus of mammalian cells, predominantly on the cytosolic surface of cis-Golgi elements (12). The two defined functions of ARF presumably involve distinct biochemical steps as well as cellular compartments, as no role for ADP-ribosylation has been described in the protein secretion pathway. While these differences are not addressed directly in this paper we have used these ARF activities to unambiguously define one functional subfamily of the GTP-binding proteins. Resultspresented in thispaper document for the first time that ARF proteins are well conserved both structurally and functionally in evolution, suggesting that the role of ARF is basic to cellular physiology and further genetic and biochemical studies inyeast will likely prove directly relevant to similar studies in man. MATERIALS AND METHODS
Cloning of Human ARFl cDNA The human ARFl cDNA was cloned by hybridization of a radiolabeled bovine ARFl cDNA probe to a human foreskin fibroblast cDNA library (13). The pCD2 library was constructed by and was the generous gift of H. Okayama (Laboratory of Cell Biology, National Institute of Mental Health, Bethesda, MD). Approximately 450,000 colonies were screened with the 700 bp PstI-PstI fragment from the bovine ARFl cDNA, labeled to 10' cpm/pg by nick translation, as described in Sewell and Kahn (8). Six potential positives were identified in the initial screen. Partial DNA sequencing was performed on the four largest inserts after subcloning into pUC19. Three of these were found to have regions of at least 100 bp which were identical and therefore assumed to derive from the same message. The largest of these was completely sequenced and used as thesource of human ARFl cDNA. Both strands of each of the human ARF cDNA inserts were sequenced by the dideoxy chaintermination method of Sanger (14). Cloning of Human ARF4 cDNA Construction of the cDNA Expression Vector-The details of the cDNA expression vector constructionare described elsewhere.' Briefly, the system utilized employs two different plasmids to maintain the cDNA expression vector in an episomal state in cells that have received both plasmids, thus allowing for the efficient retrieval of plasmids that are capable of conferring the phenotype that is being selected. The first plasmid (p266CH2) confers resistance to the drug hygromycin B (3) and is also an expression vector for Epstein Barr virus nuclear antigen 1. The second plasmid (pCNCEB8) contains two transcription units driven by the cytomegalovirus immediate early gene promoter-enhancer. The first unit confers resistance to the drug G418 in mammalian cells and kanamycin in bacterial cells (15, 16). The second transcription unit directs the expression of the cDNA. The pCNCEB8 plasmid also contains the Epstein Barr virus oriP sequences and is maintained as an episome in cells that are producing EBNA-1 (17, 18). cDNA Library Construction-The estrogen receptor-negative human breast cancer cell line MDA-MB 231 was used as the source of the mRNA used in the synthesis of the cDNA that was inserted into the pCNCEB8 expression vector. This cell line clones with high
'F. G. Kern and E. P. Gelmann, manuscript in preparation.
Factors
2607
efficiency in soft agar and also secretes a protein with fibroblast growth factor-like activity into the medium conditioned by these cells. Total cellular RNA was isolated by the method of Chirgwin et al. (19). Poly(A)+RNA was isolated and cDNA was prepared using the method of Gubler and Hoffmann (20). The cDNA wasmethylated with EcoRI methylase and EcoRI linkers were added. The cDNA was then ligated into theEcoRI site of the pCNCEB8 vector. The ligation mix wasused to transform DH5a cells and >loKcolonies were pooled and plasmid DNA was extracted and purified by double banding in CsC1-ethidium bromide gradients. Forty pgof plasmid DNAwere used to transfect an Epstein Barr virus nuclear antigen 1production clone of SW13 cells named 1SH4 cells using the procedure of Chen and Okayama (21). Two days after transfection of lo6 1SH4 cells, the cells were plated in 1.5 ml of improved modifiedEagle'smedium containing 0.35% agar and both 50 pg/ml of hygromycin B (Behring Diagnostics) and 400 pg/ml of G418 (Geneticin, GIBCO) at a density of lo5 cells/60-mm dish on top of a 4.0-ml base layer of improved modified Eagle's medium containing 0.6% agar. Northern Blot Analysis-Twenty pg of total cellular RNA, isolated in guanidium isothiocyanate, was electrophoresed in a 1.2% agarose gel containing 2.2 M formaldehyde and transferred to nitrocellulose using 20 X SSC (1 X SSC = 0.15 M NaCl, 0.015 M sodium citrate). The filters were baked ina vacuum oven at 80 "C for 2 hand prehybridized a t 55 "C for 4 h in asolution containing 50% formamide, 5 mM NaPO,, pH 6.5, 5 X SSC, 0.1% SDS, 1 mM EDTA, 2.5 X Denhardt's solution (1 X Denhardt's solution = 0.02% bovine serum albumin, 0.02% polyvinylpyrrolidone, 0.02% Ficoll), and 200 yg/ml of denatured salmon sperm DNA. The filters were hybridized In the same solution containing IO6 dpm/ml of anin vitro transcribed antisense RNA ARFl or ARF4 probe, radiolabeled with 32P,at 55 "C overnight (22). The ARFl probe consisted of a 210-bp SacI-Hind111 fragment from the coding region in the vector Bluescript SK M13+ (Stratagene) that was linearized with XhoI and transcribed using T, polymerase as recommended by Promega. The ARF4 probe consisted of a 259-bp EcoRI-EcoRV fragment from the coding region in the pGEM7zf+ vector (Promega) that was linearized with XbaI and transcribed with SP6 polymerase. After hybridization, the filters were washed with three changes of 0.1 X SSC, 0.1% SDS at 65 "C for a total of 1h. The filters were exposed to Kodak XAR-2 film overnight using Du Pont Quanta 111 intensifying screens. The filters were then stripped of radioactive signal by being placed in 0.1 X SSC at 95 "C for 30 min. After prehybridization for 4 h a t 42 "C in a solution containing 50% formamide, 5 X SSC, 0.05 M sodium PO,, 5 X Denhardt's solution, 1.0% glycine, and 100 pg/ml denatured salmon sperm DNA, the filters were hybridized with lo6 dpm/ml of a nicktranslated 1.3-kb PstI fragment containing the rat glyceraldehyde phosphate dehydrogenase cDNA in a solution of 50% formamide, 5 X SSC, 0.01 M sodium PO,, 1 X Denhardt's solution, 0.3% SDS, 10% dextran sulfate, 100 pg/ml denatured salmon sperm DNA at 42 "C overnight. The filters were washed in 2 X SSC, 0.1% SDS at 42 "C for 30 min followed by two successive washes with 0.2 X SSC, 0.1% SDS, and 0.1 X SSC, 0.1% SDS at 50 "C for 30 min each. The filters were then exposed to film overnight. Expression of Human ARFProteins in Bacteria-The purification of recombinant bovine ARFl has been described (23). As the bovine and human ARFl proteinsare 100% identical, the recombinant bovine ARFlp was used in these studies and is referred to as mammalian ARF (mARFlp). Expression of hARF4p was achieved through the use of the same vector, pET3C (24, 25), as that used previously for bovine ARF1. Unique NdeI and BarnHI sites were inserted into the hARF4 cDNA at theinitiating methionine and 6 bp downstream of the stop codon, respectively, using synthetic primers in a polymerase chain reaction. The NdeI-BamHI fragment was then ligated into pET3C, previously digested with NdeI and BarnHI. The resultant plasmid, pJCH2-1, was used to transfect competent BL21 (DE3) cells to ampicillin resistance. The growth, induction, and lysis of bacteria as well as thepurification of the recombinant hARF4p was performed as previously described for bovine ARFl (23). The pET3C expression plasmid and BL21 (DE3) cells were the generous gift of Dr. F. W. Studier (Brookhaven National Laboratories, Upton, NY). Purification entailed sequential ion exchange (DEAE-Sephacel) and gel filtration (Ultrogel AcA54) chromatography. The purified protein appears very similar to bARFlp in purity (85-90% by densitometry of Coomassie Blue-stained gels), electrophoretic mobility, and activities (see below). Expression of Human ARF in Yeast-The parent cell line for these studies (see Table 11) was PSY315 (MATO!his3-~001eu2-3,1121ys2801 urd-52). Disruptions at each ARF gene to yield TT104 (MATa
2608
Human ADP-Ribosylation Factors
his3-A200 leu23,I 12 lys2-801 urd-52 arfl::HIS3) or TT139 (MATa ade2-I01 his3-A200 leu2-3,112 lys2-801 urd-52 arf2::LEU2) were performed asdescribed in Stearns et al. (7). Plasmidsfor the inducible expression of human ARF cDNAsusing the GAL1-GAL10 promoter in yeast were based on the pBM150 vector of Johnston et al. (26) which is a low (1-2 copies/cell) copy number CEN plasmid carrying the selectable URA3 marker. The parent plasmidfor these constructions, pJCY1-38, was produced by inserting an 1100-bp fragment, including the yARFl coding region and 600 bp downstream, into the BarnHI and Sal1 sites of the pBM272(a derivative of pBM150) vector, generously provided by Mark Johnston (Washington University, St. Louis, MO). This puts the ARF coding region immediately downstream of the GAL1 promoter and provides the necessary transcription terminator(s) from the yARFl gene. The yARFl open reading frame in pJCY1-38 was then switched with either thebovine ARFl (to produce pJCB1-21) or human ARF4 (to produce pJCH2-8) open reading frames by inserting a BarnHI site immediately upstream and a XbaI site 6 bp downstream of the human ARFcoding regions using synthetic primers and the appropriate cDNAs as template a polymin erase chain reaction. This leaves the upstream (GAL1 promoter) and of human downstream regulatory elementsunaltered.Induction ARFp expression was achieved in liquid culture or agar plates by the inclusion of 2% galactose as the sole carbon source. A parallel set of plasmids was constructedallowing the constitutive expression of yARFl (pJCY1-31), mammalian ARFl (pJCB1-19), and human ARF4 (pJCH2-7) proteins using the yARFl promoterby replacing the 685-bp EcoRI-BarnHI GAL1-GAL10 promoter with the yARFl promoter ( ~ 4 0 0 - bEcoRI-BarnHI ~ region immediately upstream of the yARFlcoding region).The pJCY1-31 plasmid hasbeen shown (7) to fully complement the arfl null phenotypes and restore pARFIp to control levels. Methods for yeast growth, mating, sporulation, and tetrad analysis were performed as described in Sherman et al. (27). Yeast cells were transformed with DNA by the lithium acetate method of Ito et al. (28). Miscellaneous-Nucleotide binding and GTP hydrolysis were determined as previously described (5, 23). Assays were performed a t 30 "C in 20 mM HEPES, pH 7.5, 100 mM NaCI, 1 mM dithiothreitol, 1 mM EDTA, 0.5 mM MgCI,, 3 mM L-cr-dimyristoylphosphatidylcholine, 0.1% sodium cholate, and the appropriate nucleotide. The free magnesium concentration is calculated to be 1-10 nM under these conditions. The in vitro ARF assay was performed as described in Weiss et al. (23) and measures the ability of a protein to serve as the cofactorinthe cholera toxin-dependentADP-ribosylation of the stimulatory regulatory subunit of adenylate cyclase, Gscr. The incorporation of a radiolabel from [cx-~*P]NAD into purified recombinant Gsu (generously provided by M. Graziano, Merck and Co., Rahway, NJ) was determined bothby the filter trapping method (29) and after electrophoresis of samples on 13% polyacrylamide SDS gels (3). For both the ARF assay and GTPase assay, binding of GTPyS to ARF was determined a t identical time points and under identical reaction conditions. The activities in each assay were then normalized to the number of nucleotide-binding sites in the ARF sample to determine specificactivities. Proteins were determined using the method of Schaffner and Weissman (30). The production of ARF antibodies and immunoblotting procedures were performed as described in Kahn et ai. (6).Thedescription of themethodsusedtoquantitate noncovalently associated nucleotides to purified recombinant proteins appears in Weiss et al. (23) and includes heat denaturation of the protein followed by quantitative high performanceliquid chromatog[crraphy analysis of released nucleotides. [LU-~'P]GDP, [y-32P]GTP, '"PP]NAD, and GTPyS were purchased from Du Pont-New England Nuclear.
nously. SW13 cells transfectedwithanexpression vector directing theexpression of a secreted member of the fibroblast growth factor family plate with high efficiency in soft agar (32). However, results of experiments attempting to prove a directrelationship between ARFprotein expression and growth factor secretion are currently inconclusive. The cloning of hARF4 by this screen is perhaps bestviewed as serendipitous at this time. It is presented because it explains the origin of the clone and is consistent with the recently described role of ARF in protein secretion(12). Approximately 3 weeks after plating, four macroscopic colonies were found in the plates containing cells transfected with the MDA-MB 231 library, whereas no visible colonies were found in the plates containingcells transfected with the vector alone. One of these colonies, named SHMCDl, was successfully expanded in media containing both hygromycin and G418 and found to contain at least four different plasmids. Three of the four plasmids obtained from SHMCDl were analyzed further. The three plasmids had insert sizes of approximately 1500,950, and 2600 base pairs and were named pMCD1-1, pMCD1-2, and pMCD1-3, respectively. Retransfection of 1SH4 cells with eachof these plasmidsindividually or with a poolof all three plasmids failed to yield large colonies in soft agar. The insert present inpMCD1-1 was sequenced and found to contain an open reading frame of 540 bp which encodes a protein of approximately 2 1 kDa(see Fig. 1). A computer search of the deduced protein sequence revealed homology with members of the family of guanine nucleotide-binding proteinswith significantly greatersimilaritytotheADPribosylation factor subfamily. The pMCD1-1 cDNA insert was namedhuman ARF4 as two ARFcDNAs havebeen previously described (hARF1 and hARF3) and a human homologue of bovine ARF2 may ultimately be found. The alignment of the deduced amino acid sequence of hARF4 with previouslysequenced ARFproteinsis shown in Fig. 2. 72% hARF4p is81% identical tobovine or human ARFlp and identical to yeast ARFs. The relationship to other mammalian 88% if conservative substitutionsare ARFsincreasesto scored. The hARF4p contains the consensus sequences both for GTP-binding proteins andfor amino-terminal myristylation (33). Mammalian ARFsshown in Fig. 2 have a divergent amino terminus (residues 1-17) followed by a veryhighly conserved region (18-99), with onlythree differencesin hARF4, compared with the other mammalian proteins. Only conservative differences are observed in the yeast proteins in this highly conserved region. Interestingly, this is the same region (between the first andsecond consensus GTP-binding domains) which is conserved in p21 ras proteins and thought to be involved in binding to theeffector. Cloning and Sequence Analysis of Human ARFl -The partial sequences of human ARFl cDNAs have been published previously (10, 11). In an effort to obtain a full-length copy RESULTS of the human ARFl cDNA, an Okayama-Berg library was Isolation of the Human ARF4 cDNA-The human ARF4 screened by cross-species hybridization usingthe radiolabeled cDNA was isolated as the result of an attempt to use the coding region of the bARFl cDNA as a probe, as described under "Materials and Methods." The restriction sites and approach of transfection with cDNA expression libraries to isolate genes capableof conferring an anchorage-independent cDNA and derived protein sequences of the human ARFl are phenotype to SW13 cells, derived from an adrenal cortical shown in Fig. 1. The naming of human ARFl is based on the carcinoma cell line (31). Despite the presenceof micromolar 100% agreement of the human and bovine ARFl (8) protein previously reported (10, 11). The concentrations of intracellular basic fibroblast growth factor sequences, ashasbeen (bFGF) (32), these cells do not secrete detectable levels of nucleotidesequence of the human ARFl cDNA is also in cDNA sequence previously bFGF and do not formcolonies in soft agar unlesspicomolar perfect agreement with the partial concentrations of growth factors that are members of the published (10, 11),although nearly a kilobase more of the 3'fibroblast growth factor family of peptides are added exoge- untranslated region is shown in Fig. 1. This allowed a com-
Human ADP-Ribosylation Factors Na
PH
pi
HzN"~
'~
o
N ~
~~~
0,6
0,4
.~
~~
-~~ 'COOH
fi,RFl
~
02
0-8
lL2"
lJ0
1,6
,?4
15kb
'
;
G
R
A
h
A
C
C
A
A
C
G
C
C
T
G
G
C
T
C
t
G
R
G
~
A
~
C
~
~
~
~
~
~
~
~
bARF 1 : bARF2 : (,(,GAACRTCTTCGCCARCCTCT~CARG~GCCTTTTTGGCCTTT~T~~GC-AAAAGAAATGCGCATCCTCATGGTGGGCCTGGATGCTGC hARFl : G N I F A N L ~ K G L F G K K E M R I L M V G L D A A hARF3 : ~:;r~~AaGRCCACGRTCCTcTAcAAGCTTAAGCTGGGTGAGATCGTGACCACCATTCCCACCATAGGCTT~~~~~Gh ~ ~ ~m~ 4 : S K T T I L Y K L K L G E I V T T I P T I G F N V E yARF1: yARF2 : cC6TGGRGTACAAGAACITCAGCTTCACTGTGTGGGACGTGGGTGGCCaGGAC~GATCCGGCCCCTGTGGCGCCACTAC
Y
MGNIFGNLLKSLIGKKEMRILMVGLDAAGKTTILYKLKLGEIVTTIPTIG MGLTISSLFSRLFGKKQMRILMVGLDAAGKTTILYKLKLGEIVTTIPTIG MGLFASKLFSNLFGNKEMRILMVGLDGAGKTTVLYKLKLGEVITTIPTIG
so so so so so so so
MGLYASKLFSNLFGNKEMRILMVGLDGAGKTTVLYKLKLGEVITTIP~IG
K*MRILMVGLD*AGKTT*LYKLKLGE**TTIPTG
80 ;
M
T
MGNIFANLFKGLFGKKEMRILMVGLDAAGKTTILYKLKLGEIVTTIPTIG
GLMG L
B ~
MGNIFANLFKGLFGKKEMRILMVGLDAAGKTTILYKLKLGEIVTTIPTIG MGNVFEKLFKSLFGKKEMRILMVGLDAAGKTTILYKLKLGEIVTTIPTIG
bARFl : b-2 : hARF 1 : hAW3: hARF4 : yARF 1 : yARF2 :
A
Pv
2609
E
Y
K
N
I
S
F
T
V
W
D
V
G
G
Q
D
K
I
R
P
L
W
U
H
Y
~
~
~
~
~
~
~
~
~
~
~
G
~
~
~
~
~
~
~
~
~
T
~
~
~
~
~
FNVETVEYKNISFTVWDVGGQDKIRPLWRHYFQNTQGLIFWDSNDRERV FNVETVEYKNISFTVWDVGGQDKIRPLWFLYYFQNTQGLIFWDSNDRERV
~
~
~
~
~
100 100
FNVETVEYKNISFTVWDVGGQDKIRPLWRHYFQNTQGLIFWDSNDRERV FNVETVEYKNISFTVWDVGGQDKIRPLWRHYFQNTQGLIFWDSNDRERV FNVETVEYKNICFTVWDVGGQDRIRPLWKHYFQNTQGLIFVVDSNDRERI FNVETVQYKNISFTVWDVGGQDRIRSLWRHYYRNTEGVIFWDSNDRSRI FNVETVQYKNISFTVWDVGGQDRIRSLWRHYYRNTEGVIFVIDSNDRSRI FNVETV'YKNI FTVWDVGGQD*IR*LW*HY* NT*G*IFV*DSNDR R *
100
150
100 100
100 100
?TCCAGAACRCACAAGGCCTGATCGTGGTGGTGGACAGCAATGACAGAGAGCGTGTGRACGAGGCCCGTGAGGAGCTCAT
RTCCGGCCGRGATCRCAGACRI\GCTGGGGCTGCACTCACCACT~CGCCACAGG~ACTGGTACA~TC~GGCCACCTGCGCCACC
bARFl : bARF2 : hARF 1 : hARF 3 : hARF 4 :
N
yARFl :
NEAREELMRMLAEDELRDAVLLVFANKQDLPNAMNAAEITDKLGLHSLRH NEAREELTRMLAEDELRDAVLLVFVNKQDLPNAMNPAEITDKLGLHSLRQ NEAREELMRMLAEDELRDAVLLVFANKQDLPNAMNAAEITDKLGLHSLRH NEAREELMRMLAEDELRDAVLLVFANKQDLPNAMNAAEITDKLGLHSLRH QEVADELQKMLLVDELRDAVLLLFANKQDLPNAMAISEMTDKLGLQSLRN GEAREVMQRMLNEDELRNAAWLVFANKQDLPEAMSAAEITEKLGLHSIRN
yARF2 :
GEAREVMQRMLNEDELRNAVWLVFANKQDLPEAMSAAEITEKLGLHSIRN
F
O
N
T
Q
G
L
I
F
Y
V
D
S
N
D
R
E
R
V
N
E
A
R
E
E
L
M
( j R ( ; I ; A T G C T G G C C G A G C A C G R G C T C C G G G * T G C T G T C C C C ~ A C G C C A T G ~ K M L A E D E L R D A V L L V F A N K Q D L P N A M
R
A
E
I
T
D
K
L
G
L
H
S
L
R
I
I
R
N
W
Y
I
Q
A
T
C
A
?
nC.C'j~CGnCGGGCTCTATGARGGACTtiGACTGGCiCTCCAATCAGCTCCGGRACCAGAAGTGAACGCGACCCCCCTCCCT S C D G L Y E G L D W L S N Q L R N Q K
E
* * *ML DELR*A L*F
NKQDLP'AM*E*T'KLGL
150 150 150 150
150 150
S'R
CTCRI:TCCTCTTGCCCTCTGCTTTRCTCTCTCA~GT~GCAAACGTGCGGCTCG~GGTGTGAGTGCCAG-GC~GCCTCCGTG
bARF 1 : bARF2 : hARFl : hARF3 : hARF 4 : y ARF 1 : yARF2 :
~TTT~STCRCCGTGTGCATCGCACCGTGC~GTIRR~GTGGCAGACGCAGCCTGCGGCCAGGCTT?TTATTTAATGTAAAT AS~T~TTGTTTCCRATGA~GCAGTTTC~GGTACTCCTATGCAATATTACTCAGCTTTTTTTATTGTAAAAAGAA-ATCA ACrCACTGTTCAGTGCTGAGAGGGG~TGTAGGCCCATGGGCACC~GGCCTCCAGGAGTCGCTGTGTTGGGAGAGCCGGCC ACGCCCTTGGCTTAGRGCTGTGTTGRRATCCATTTTGGTGGTTGGTTTTAACCCAAACTCAGTGCATTTTTTAAAATAGT 7 9 L l i l , n T I : C A A t T C G A G A A C I C T T G A A C A S A C I I G I A C
RNWYIQATCATSGDGLYEGLDWLSNQLRNQK
RNWYIQATCATSGDGLYEGLDWLSNQLKNQK RNWYIQATCATSGDGLYEGLDWLSNQLRNQK RNWYIQATCATSGDGLYEGLDWLANQLKNKK RTWYVQATCATQGTGLYEGLDWLSNEL SKR RPWFIQATCATSGEGLYEGLEWLSNSLKNST RPWFIQSTCATSGEGLYEGLEWLSNNLKNQS R W"Q*TCAT G GLYEGL'WL*N L*
181 181 181 181 180 181 181
FIG. 2. Alignment of bovine, human, and yeastARF protein sequences. The cDNA derived amino acid sequencesof ARF proteins are shown (see text for references of previously published sequences). ?li~~~~GGGCCAGGC~CGGAGCCCACA~CC~CC~C~C~,,TGCCGC~GGCCGCCCTAC~CACCTTCAGG~AGCCTATtiGGA Those residues conserved in all ARF proteins are shown on the last CGC-(,GCCCCRTCTGTCCCTCGCiTCCGCGTGTGtiCCAGAGTGGTCCtiTCGTCCCCAACACTCGTGCTCGC~CAGACAC'rT line of each grouping and those with only conservative substitutions TDGCI\GGaTGTCTGGGGCC~CACCAGCRGCGCGTGCGCGTGCAAtiCCGGGCAGGCGGTCC~CCTAGACCCACAGCCCCTCGGG are indicated by an asterisk.
CA~~CGCCRGCCCAGC?GTTCCCCTCGGGGAACATGAGGTGGTGGTGGCGCAGCAGACTGCGATCAATTCTGCATGGTCAC
nili',AGaTCCCCGCARCTCGCTTGTCCTTGGGTCACC~'rGCATTCCA?AGCCATGT~CTTGTCCCTGTGCTCCCACGG~
a~~ilCCCACCTCTGTGTGTGATG~AGCTTTCTCTCCCTCA~~CTGCAAG~GTCCGATTTGCCATCGAAAAAGACAACCT CTIICT:TTTTCTTTTGTATTTTGA~~~ACACTLRR:CTl;tiA[;CrGTTAAArTTATCTTGGGGAAA~CTCAGAhCTGGTC'r
I
II
Ill
IV
hlTliG-GTCGTAGGRACCTCTTRCTGCTTTCAnTRChCGnTTATTGTATACTTGTTTTCAGTTTTC
5'
hf1:CGACRAACAAGCAiTGT-TTATRGCTAT~~GAATAA~.ATCTCTTAACTATT~AAAA~l825
C
Nc
RI P
RV
T
Rs
FIG. 3. Alignment of human ARFl and bovine ARFl cDNA. Alignment of the two cDNA sequences was performed with the aid 0 0.2 0.4 1.01.4kb1.2 of the ALIGN program of the Microgeniesoftwarepackage. The overall alignment scoringof the cDNAswas 74.5% identity with 1408 D matches, 267 mismatches, and216 unmatched (108 occurring a t ends due to differences in overall length) out of a total length of 1,891 80 CGRGRAAGCACGGGGTCGCCCCAAACCCCTTCTGCTTCTGCCCATCAC-G~GCC~C~*CCGCCATGGG~CTCACTATCT nucleotides. Those areas with scores of 88% and above are indicated M G L T I by a roman numeral, with I being the coding regions of the ARF CCTCCCTCTTCTCCCGACTATTTGGCAAGAAGCAGATGCGCATTTTG~TGG~TGG~~~GGA~GCTGC=GGc~GAcRAC~ ~ S L F S R L F G K K P M R I L M Y G L V A A G K ~proteins. ~ I
#
1
I poly A
COOH 0.6 0.8
ARF4
I
ATTCTGTATRRACTGRAGTTAGGGGAGATAGTCACCACCATTCCTACCATTGGTTTTRATGTGGA-CAGTAGAA~~~I
~
~
Y
K
L
K
L
G
E
I
V
T
T
I
P
T
I
G
F
N
V
E
T
V
E
~
K
parison of the untranslatedregions of bovine and human ARF cDNAs. A~GGTCTTATTTTTGTGGTAGATAGCRACGATCGTGARAGRA~TCAGGRAGTAGCAGATGAGCTGCAG-TGCTTCTG P G L ~ F V V D S N D R E R I Q E V A D ~ ~ Q K ~ Alignment L L of the nucleotide sequences of bovine ARFl and G ~ A G A ~ G A A T T ~ A G A G A T G C ~ G T G C T G C T A C T T T T T G C ~ C A A A C A G G A T T T G ~ C - T G C T A T G G C C A T C A ~ ~ ~human ~~~~ coding ARFl cDNAs reveals 90.4% identity within the ~ D E L R O A V L L L F A N K Q D L ~ N A ~ A ~ ~ ~ M regions with no gaps. A search of additional homologous GACAGATAARCTAGGGCTTCAGTCTCTTCGTAACAGAACATG'TATG~~C*~GCC~C~~G~GCRAC*C~*GG~~~~~~~~ T D K L G L Q S L R N R T W ~ V Q A T C A T Q G T ~ regions between these two cDNAs revealed other regions, in T G T A T G n n G G A C T T G A C T G G C T G T C A A A T G A G C ~ ~ ~ C ~ ~ C ~ = ~ * - ~ ~ - ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ the untranslated regions, with a degree of relatedness comL Y E G L D W L S N E L S K R A T A A A A T T G G T C T R G G C T T G T T A C A A C A A A a T T A G T T T G T A T c T ~ G G ~ ~ ~ ~ ~ A ~ ~ c A G T A T c T G G G ~ c T G G T T T G G G c A Gparable to that seen in the coding region. These regions are indicatedas ZZ, IZZ, and ZV in Fig.3. Shorter regions of AATA'ITAAACTTATTTTGTTGCCAATTATTGT~~ACCGAGTATAATGTTGCTAT,~~AGCRATGTGCTTGGTTTT-~~ near the coding region of bovine and human A ' T ~ C T C C T ~ G G G A A A A A d G T A T C C T C T T T T A R T T T T A A T T T T A C T T C C C A T A A G C G T A A A T G c c T G G A c ~ T A ~ ~ T ~ T ~ t i T G A ~conservation ~~T have been noted previously (10). In contrast, no other T T A A A T A A A T T G T T T G A ~ T G T T T T T G A G C C C C A G A C A R R T G T ? ~ ~ C C C T T G C T ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ARFl ~,~ T * ' l C A r T C C T G A G A C A G T T T G C T - T T T a A A a ~ ~ G r A G c * ~ ~ c C * ~ ~ ~ G ~ * ~ T T ~ ~ ~ ~ c ~ c T C c c T ~ G C c A A A A Aregions GATT~ of homology were detected in the untranslated regions T ~ T ~ A ~ ~ ~ ~ ~ ~ T T G T A C C R G C C A G A G A A A G A T C C A A A A c A c ~ A c ~ c A G C ~ c ~ ~ ~ of c ~ ~the ~ ~ other ~ ~ ~ ~ ~yeast, ~ ~ ~ ~ human, ~ ~ ~ ~ ~ c cor T ~ c ofc ~ any bovine ARF DNA A T T G A C T C C T G G C C T A C A T C A G C C R A R C T T A A C T T A A C ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~sequences. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ T T ~ T ~ T ~ ~ T G G T T G c A a A C R A T T G A T A T T T A G A T G G T T l T T l i R T R C T C A G C A G ~ r T G T C T ~ c C c ~ ~ ~ T T G ~ G ~ c ~ ~ ~ T T ~ A T ~ T T ~ c A ~ Northern ~TTGcT Blot Analysis of H u m a n A R F land ARF4 ExpresT T T G T T R T C A G C C T G A T T T T T T G C T C ~ G ~ ~ ~ ~ ~ t i ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ G A c A T A T c T T c A T T A sion-Northern blotting using human ARFl as probe under A G A G T T T T T G G R A A A C T C A T C A A ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ T ~ ~ ~ T T A T T C C T R A T - T G A T A stringent conditions of washing have previously identified a AAATACGIARARRAAAA-1457 single message at ~2.2-2.3 kb(10) or -1.8 kb (11).At lower FIG. 1. Nucleotide and deduced amino acid sequences and stringency an additional band a t 3.7 kb was observedby Bobak restriction maps of human ARFl and human ARF4 cDNAs. et al. (11). When a human ARFB probe was used a strong The restriction maps of human ARFl (A) and human ARFB ( C ) are shown with restriction sites indicated, using the following abbrevia- signal was observed at 3.7 kb and a weaker signal migrating a t 1.2 kb was observed (11).It is not clear if the 1.2 and 3.7 tions: P, PstI; H, HindIII; Na, NarI; N , NcoI; Pu, PuuII; A, AccI; RI, EcoRI; R V, EcoRV; Rs, RsaI; T, TaqI. Nucleotide and deduced amino kb messages represent products of related but distinct ARF genes or alternatively, derive from alternative splicing of a acid sequences of hARFl ( B ) and hARF4 (D)are also shown. single ARF gene. The results of the Northern analysisshown GAACATTTGTTTCACAGTATGGGATGTTGGTGGTC~GATAG~~~~GGCC~CTC~GGAAGCATTACTTCCAGRA~A~CC N I C F T V W V V G G Q D R I R P L W K H = ~ Q N T
~
~
~
-
Human ADP-Ribosylution Factors
2610
in Fig. 4 indicate that thepredominant band which hybridizes appears that the cDNA sequence of hARFl and hARF4 rewith an ARFl probe in these studies has anestimated size of ported here are complete, or very nearly so. approximately 2.1 kilobases. Two additional bands at 4.1 and The Northern analyses also indicated that the cell line 1.4 kilobases are also observed (Fig. 4A). It is likely that the SHMCD1, which contains episomal copies of a ARF4 cDNA 1.8- and 3.7-kb species described in Bobak et al. (11) are the expression vector, does not significantly overexpress the same asthe 2.1- and 4.1-kb species seenin Fig. 4. The mRNA for this gene when compared with the levels observed differences in estimated message size are likely due to differ- in either the parental 1SH4 or SW13 cells. Similar results ences in the gel systems used and thedegree of gel resolution. were obtained when clones of 1SH4 cells transfected solely The ARF4 probe hybridizes to a single message of about 1.7 with the pMCD-1 (ARF4) expression vector were analyzed kilobases (Fig. 4B). Despite the 71% identity at thenucleotide (data notshown). The same vector has been used to successlevel between the ARFlprobe and thecorresponding area of fully overexpress the cDNA for the kfgflhst oncogene in1SH4 the ARFl cDNA and the 64% identity between the ARF4 cells, suggesting that theremay betight control onthe amount probe and the corresponding area of the ARFl cDNA, there of ARFl mRNA that can be expressed by these cells. No was no cross-hybridization observed between ARFl andARF4 antibodies are available that are capable of discriminating under the conditions used in these studies. The 1.4-kb species between the threeknown human ARF proteins so the relative appears to be differentially expressed in the different cell lines contributions of these three genes to the total cellular ARF and tissues examined here, being most highly expressed in protein is unknown. the adrenal cortical carcinoma cell line and least expressed in Biochemical Analysis of Recombinant Human ARF Prothe placenta and estrogen-independent breast cancer cell line. teins-As described above, human ARF4 is the most divergent The additional band at 1.4 kb detected with the hARFlprobe of the five mammalian ARFs cloned to date. An analysis of may be related to the 1.2-kb band observed by Bobak et al. the biochemical and in vivo functions of human ARF was (11)with an ARF3 cDNA probe. Although they reported no undertaken in an effort to determine if human ARF4 is a hybridization of an ARFlprobe to thisspecies, our ability to bona fide member of the ARF subfamily of 20-kDa GTPdetect this species may be due to the increased sensitivity of binding proteins as defined both functionally and structurally. the RNA probe used in these studies. Alternatively, this may It has recently been documented that despite the absence of also reflect tissue-specific differences in gene expression be- the amino-terminal myristic acid, recombinant bovine ARFl tween the human brain tissue used in their experiments and retains all of the in vitro biochemical characteristics of ARF the tissue and cell lines used inour experiments. Other purified from mammalian tissues, including nucleotide bindexplanations for the observed band at 1.4 kb include hybrid- ing requirements for lipids, detergent, and magnesium, and ization to another closely related member of the ARF gene activities in both GTPase andARF assays (23). family or alternatively spliced forms of ARFl or ARF3. The Human ARF4p was expressed at high levels and purified 1.4-kb species observed with the ARFl probe and the 1.7-kb from bacteria as described under "Materials and Methods." band observed with the ARF4 probe are clearly distinct as the The level of expression, estimated at 5-10% of total bacterial former migrates below the band hybridizing to the glyceral- cell protein, and purification protocol were indistinguishable dehyde phosphate dehydrogenase probe while the latter mi- from those previously reported for mARFlp. The purity of grate above it. Allowing for the addition of a poly(A) tail it the recombinant protein was estimated to be at least 88%by laser densitometry of Coomassie Blue-stained SDS gels. The purified protein is at least 70% active as determined by A B quantitation of non-covalently associated nucleotide (GDP) which copurifies with the recombinant protein (23). 1 2 3 4 5 1 2 3 4 5 6 Although the recombinant hARF4p was soluble in bacterial 9.49 - 9.49 lysates and purified in the absence of any added lipids or 7.46 - 7.46 detergents the ability to bind activating guanine nucleotide 4.40 - 4.40 triphosphates (e.g. GTP or GTPyS) was observed only in the 4.1 presence of added lipid and cholic acid. This near absolute 2.1 237 2.37 dependence on ahydrophobic environment hasbeen described 1.7 - . I I 1.4 135 previously for purified bovine brain and bacterially expressed - 135 mARFlp and is a unique feature of ARF among the many regulatory GTP-binding proteins. As the rate of release of tightly bound GDP was previously shown to be rate limiting -0.25 - 0.25 in the nucleotide exchange reaction, the off rates of GDP from mARFlp andhARF4p were compared and also found to be very similar (tH= 23 and 20 min, respectively at 30 "C). Members of the superfamily of GTP-binding proteins charFIG. 4. Northern blot analysis of human arfl and arf2 acteristically have low but detectable intrinsic GTPase acexpression. Twenty pg of total RNA was separatedin a 1.2% tivities (typically 0.1-1.0 min"). The ARF proteins reported agarose, 2.2 M formaldehyde gel and hybridized to P-32 labeled to date (including purified bovine brain and recombinant antisense arjl ( A ) and arj2 ( B ) RNAprobes as described under "Materials and Methods." Total RNA from lane 1 , human placental mARFlp) have no detectable GTP hydrolytic activity (23). tissue; lane 2, MDA MB231 cells; lune 3, MCF-7 human mammary We were unable to detect any significant GTPase activity carcinoma cells; lune 4 , SHMCDl cells; lane 5, 1SH4 cells; lune 6, over background with either purified mARFlp or hARF4p parental SW13 srlrmnl m r t i r a l oaroinnmn aollo. Numboro to t l u ri& (=ZO.OZ p111u1uf GTP nytirolyzeCl/pmol of G'l'P-bmding site/ of the gel lanes indicate the positions of molecular size markers in min) in these studies. kilobases. Numbers to the kjt refer to the calculated sizes of the ARF Activity of hARF4p"The ability of hARF4p to serve hybridizing bands.GAPDHrefers to the position of the bandobserved when the filters were rehybridized to a rat glyceraldehyde phosphate as cofactor in the cholera toxin-dependent ADP-ribosylation dehydrogenase nick-translated cDNA probe as a control for the of Gsa was also determined. This is a sensitive and highly loading of RNA in all lanes. specific assay for ARF proteins, allowing detection of subpi-
"
---
Human ADP-Ribosylation Factors
2611
ARF Gene in Yeast-The biochemical data presented above, in particular the activity in the ARF assay, provide evidence of strong functional conservation among the ARF.proteins from different species. However,the significance of the cofactor activity in the cholera toxin reaction (ARF assay) as an indicator of the physiological function of ARF is unproven. For this reason, a more stringent, in uiuo test of ARF protein function was sought. The recent results of Stearns et al. (7), demonstrating that ARF is an essential gene in yeast, have provided such an opportunity. Plasmids allowing the constitutive or inducible expression of yeast or mammalian ARF proteins were constructed to allow a test of the ability of mammalian ARFs to function in yeast and rescue the lethal double arf 1-72- mutations. As overexpression of yARFlp in yeast was also found to be deleterious to yeast cells, we originally concentrated efforts on expressing mARFlp at levels comparable to those of yARFlp by constructing alow copy number (CEN) plasmid, pJCB1-19 (see Table 11), carrying the coding region of mARFl under control by the yARFl TABLEI promoter. However, maintenance of the pJCB1-19 plasmid ARF activities of recombinant mammalian and yeastARFs Results of binding and ARF assay are averages of triplicates with was insufficient to allow rescue of cells which lacked a funcless than a 10% error for each entry. Protein, G T P r S binding, and tional yeast ARF gene. This was attempted in two ways. First, ARF assays were performed as described under "Materials and Meth- cells (RT103) disrupted at the ARFl locus (arfl::HIS3) and ods." Assays were performed in parallel with the same reagents (in carrying a plasmid capable of expressing themARFlp particular diluted ARF samples, activated cholera toxin, and soniat 30 "C and stopped after (pJCB1-19) were crossed to a strain which was disrupted at cated L-a-dimyristoylphosphatidylcholine) 20 min. ARF assays contained 2.2 pmol of Gsa/point. ARF activities ARF2 (TT139, arf2::LEU2), sporulated, and more than 50 have been corrected for activity present in the absence of added ARF tetrads were dissected. The double ARF mutant spores (phe(approximately 0.03 pmol). The unit of specific activity is pmol of notypically HIS'LEU') were in all cases inviable, including ADP-ribosylated Gsalpmol of activated ARF/2O min. those which should contain the plasmid and thuswere capable Protein GTPrS ADP-Ribosylated Specific of expressing mARFlp. The second method utilized direct bound Gsa activitv disruption of the ARF2 locus (to arB::LEU2) of RT103 cells. pmol pmol pmol Again, we were unable to recover any cells whichwere mARFlp HIS'LEU'URA', suggesting that mammalian ARFl under 0.650 Exp. I 10 0.82 1.26 control of the yeast ARFl promoter could not rescue the 0.285 5 0.52 1.82 inviability resulting from loss of both chromosomal copies of 1 0.038 0.04 1.05 ARF in yeast. As a control, either method allowed the isolation Exp. I1 5 0.098 0.294 3.00 of strains which lacked both chromosomal copies of ARF hARF4p when the plasmid used contained the coding region of yARFl Exp. I 0.077 10 1.07 13.90 instead of that of the mammalian ARF1. Exp. I1 1 0.0034 0.068 20.00 In contrast, overexpression of the mARFlp by the strong 0.5 0.012 0.0007 17.14 GALl promoter did allow cells to survive in the absence of yARFlp any yeast ARF. Cells(RT121) disrupted at ARFl and carrying Exp. I 1 0.110 1.68 15.27 a plasmid (pJCB1-21) with mARFl under GALl control were Exp. I1 0.24 0.245 0.013 18.85 crossed to TT139 (arf2::LEU2), sporulated, and tetrads were comole amounts of ARF with no activity detected for other proteins including other members of the family of small GTPbinding proteins. This assay has previously been used to quantitate ARF activity from bovine brain, rabbitliver, turkey erythrocytes, bovine sarcolemma, and human placenta (2). Under the conditions of this assay, purified recombinant mARFlp and bovine brain ARF have very similar specific activities (23). Purified recombinant ARF from three different species were assayed, including mARFlp, hARF4p, and yeast (S. cereuisiae) ARFlp. The results are shown in Table I. The specific activities of hARF4p and yARFlp are very similar and approximately 4-10-fold higher than that of mARFlp. Thus, the multiple ARF proteins expressed in a single organism and in diverse species, as distant in evolution as yeast and man, each have cofactor activity in this specific in vitro assay. Human ARFs Can Satisfy the Requirement for a Functional
TABLE I1 Strains and plnsmids used in studies of ARF expression in yeast
PSY315" TT104" TT139" 121.13C RT166 RTlOl RT119 RT103 RT121 RT123 RT161
TT104 carrying pJCY1-31 TT104 carrying pJCY1-38 TT104 carrying pJCB1-19 TT104 carrying pJCB1-21 TT104 carrying pJCH2-7 TT104 carrying pJCH2-8
pBM272 pJCY1-31 pJCY1-38 pJCB1-19 pJCB1-21 bARFl pJCH2-7 pJCH2-8
derivativepBM150'; of CEN4 URA3 plasmid with GAL1-10 promoter yARFl coding region behind yARFl promoter yARFl coding region behind GALl promoter bARFl coding region behind yARFl promoter coding region behind GALl promoter hARF4 coding region behind yARFl promoter hARF4 coding region behind GALl promoter
a his3-A200 leu2-3,112lys2-801 ura3-52 Gal' a his3-A200 leu2-3,112lys2-801 ura3-52 arfl::HIS3 Gal+ a ade2-101 his3-A200 leu2-3,112 lys2-801 ura3-52 arf2::LEUa Gal+ 01
ade2-101 his3-A200 '-u2-3,112 lys2-801 ura3-52 arfl::HIS3 arf2::LEU2 (pJCB1-21) Gal'
a his3-A200 leu2-3,112lys2-801 ura3-52 arfl::HIS3 arB::LEU2 (pJCH2-8) Gal+
For details see Stearns et al. (7). and Davis (26).
* For details see Johnston
2612
ADP-Ribosylation Human
Factors
dissected and grown on plates containing complete medium the level of 0.03-0.1% of total cellular protein (not shown), and galactose as sole carbon source. Of 50 tetrads dissected with yARFlp accountingfor about 90% of total ARF protein only one spore, 121.13C, had the genotype HIS’LEU’URA’ (7). In contrast, 121.13C expresses mARFlp at thelevel of 1predictedfora cell disrupted a t both yeast ARF loci and 2% of total protein. Consistent with this result is the obsercarrying theplasmid. The 121.13c cells also failed to grow in vation that mARFlp is easily detected as a major cellular completemedia in which glucose was thecarbon source, protein on Coomassie Blue-stained gels of total cell extracts consistent with the requirement for galactose induction of prepared from 121.13C cells. This apparent requirement for mARFlp expression for cell growth. very high level expression of mARFlp may explain the low The expression of mARFlp and lack of expressed yeast frequency in obtaining viable spores lacking both yeast ARF ARFs in 121.13C were confirmed by immunoblotting using a genes. Incontrast,the level of hARF4pexpressed under combination of ARF antisera, as shown in Fig. 5 . Purified control of the GALl promoter in the absence of any yeast recombinant mARFlp (lane 2 ) and a yARFlp fusion protein ARF proteins was indistinguishablefrom that of yARFlp (lune I), containing the entirecoding region of yARFl and an under control by its own promoter (not shown). As the plasadditional 10 amino acids at the amino terminus, were used midsused to express mARFlp and hARF4p under GALl as standards. Because the ARF proteins are so highly struc- control were identical except forthe ARFcoding regions there turally conserved, the antibodies used show some degree of appearstobestrong selective pressure for strains which cross-reactivities. With the additional help of slight differ- express each of these essential proteins a t (different) levels ences in electrophoretic mobilities (mARFlp > yARF2p > optimal for cell growth. Thus, the two human ARF proteins yARFpl) it was unambiguously demonstrated that the imtested appear to have different “specific activities” in yeast, munoreactivity in 121.13C (lane 5) is the result of expressed withhARF4p similartoyARFlpandmARFlp 3-10-fold mARFlp and that no yeast ARF proteins are expressed in lower. Interestingly, this agrees very well with the specific this line. This is perhaps most easily seen with the R-40 activities observed for these proteins in the i n vitro ARF assay cYyARFlp antibody (bottompanel, Fig. 5). (see above). Expression of hARF4p was also found capable of rescuing DISCUSSION yeast cells devoid of yeast ARFs. Disruption of the ARF2 gene of RT123 cells resulted in yeast cells (RT166) lacking any As the listof small molecular weight GTP-binding proteins yeast ARF proteins but carrying a plasmid (pJCH2-8, see grows, it becomes increasingly important to be able to group “Materials and Methods”) containing the hARF4coding re- similar or like proteins together based on assays of function, gion under controlof the GALl promoter. Immunoblot analyrather thansequence information alone. This paperdescribes sis of ARF proteins expressed in RT166 confirmed the lack the use of specific in vitro and i n vivo ARF assays,each of detectable yeast ARFs and the presence of hARF4p (not involving functionallyimportant protein-protein interactions, shown). to determine the extent of functional conservation among Quantitative immunoblotting, by comparison of immuno- structurally related GTP-binding proteins. The results indireactivity seen in total cell extracts to purified recombinant cate a remarkable degree of functional conservation between proteins, revealed that yeast ARF proteins are expressed a t human, bovine, and yeast ARF proteins,including the newly identified hARF4p. The ability of these functional assays to discriminate between structurally related proteins is further emphasized by recent results‘?demonstrating the presence in D. melanogaster of an essential GTP-binding protein which is 55% identical to mammalian ARFs but has no activity in either of the ARF assaysdescribed above. Assays of fruit fly extracts confirm the presence of ARF activity in this organism, presumably the productof a currently unidentified ARF gene. Thus, the superfamily of small GTP-binding proteins, when fully enumerated, may consist of proteins with acontinuum of degrees of relatedness to any one member. These results further stress the importance of functional screens in identifying new genes in specific pathways ashARF4 was not detected by nucleic acid hybridizations with hARF1, hARF3, or bARF2 cDNAs. The specter of even more human ARF or ARF-related genes clearly cannot be ruled out. Two ARF genes with about 96% identity have previously been observed in cows, yeast, and man. The human ARF4 gene, described above, is actually the third humanARF gene cloned and analysisof the predicted protein sequence revealed that hARF4p is the most structurallydivergent of the mammalian ARF proteins. Indeed, the degree of identity between FIG.5. Immunoblots of yeast strains expressing yeast or hARFlp andhARF4p (81%) isonly slightly greaterthan that mammalian ARF1. Immunoblots from five different strains are between human ARFl and yeast ARFl (74%). Assuming that shown with purified recombinant protein controls. Proteins (30 ng) multiple ARF genes in a single organism arose as a result of and cell lysates (30 pg) wereprepared and immunoblotted as described Ilnrler “Mstnrialc and Mothodo.” Lanoo; 1, purifiod
yhnFlp fu&ll
protein; 2, purified mARFlp; 3, TT104 (arfl::HIS3 ARF2) grown in 2% glucose; 4, TT139 (ARF1 arf2::LEU2) grown in 2% glucose; 5, 121.13C (arfl::HIS3 arf2::LEU2 pJCB1-21) grown in 2% galactose; 6, RT121 (arfl::HIS3 ARF2 pJCB1-21) grown in 2% glucose; and 7, RT121 grown in 2% galactose. The antibodies used in each panel are listed a t the left and aredescribed under “Materials and Methods.”
gene duplication, it amears thathARF4 emereed auite earlv
in humanevolution. No clear reasonhas been foundto explain the needformultiple ARF proteins in a single organism although differences in the levels of expression (7) and post-
’J. W. Tamkun, R.A. Kahn, M. Kissinger, M. P. Scott, B. J. Brizuela, and J. A. Kennison, manuscript submitted.
Human ADP-Ribosylation Factors translational modification (6) have been described. Analysis of the untranslated regions of the cloned cDNAs revealed regions of conservation between human andbovine ARFl (see Fig. 3). These conserved regions are not found in other mammalian ARFs or in yeast ARF1. The occurrence of these stretches of conserved UTR segments indicates that segments of the untranslated regions of at least some ARF messages are under considerable selective pressure. The implication is that these conserved sequences may have functional roles in the regulation of the expression, stability, or targeting of specific ARF messages. No such evidence is currently available for ARF messages, although examples in whichsequences downstream of the transcriptioninitiationsite have been shown to modulate promoter activities have been reported (34, 35). It has been noted (35) that all of the previously identified messages known to contain conserved regions in the 3’ UTR code for structural proteins (actin, tubulin, or collagen). This is particularly interestingly in light of recent evidence4demonstrating that ARF is one of the coat proteins found on Golgi-derived non-clathrin coated vesicles. There is good evidence of stringent control of the levels of expression of ARF proteins in S. cerevisiae. Either deletion or overexpression of ARF in yeast resultsin loss of cell viability. Strict limits on the limits of human ARF proteins required to rescue the double arf disruption may explain the low frequency of obtaining cells lacking both chromosomal ARF genes. The rank order of potency of yARFlp = hARF4p >> mARFlp in the in vitro ARF assay is consistent with the levels of ARF proteins found expressed in yeast lacking chromosomal copies of yARFl andyARF2. The results described above strongly suggest, however, that when fully elaborated the cellular role of ARF will prove to be constant in evolution. Results presented above indicate a good correlation, both qualitatively and quantitatively, between the i n vitro and i n vivo ARF assays. This somewhat unexpected result should facilitate furtherbiochemical studies of the mechanism of action of ARF by allowingthe use of the quicker and more easily quantitated i n vitro ARF assay as an indicator of the physiological function of ARF. A role for ARF proteins in the regulation of secretion of bFGF is suggested but clearly remains unproven by the results reported here. This paper describes the first results of the analysis of ahuman mammary carcinoma cDNA library screen designed to identify one or more gene products involved in the regulation of secretion of members of the fibroblast growth factor family of growth factors. As neither the route of growth factor secretion nor the mechanism of regulation is known, it was originally unclear if the expression of a cDNA insert would allow the anchorage-independent clonal growth of the human adrenal carcinoma cells. Unfortunately, these results come from a single isolate and the inability to reproduce anchorage independent cell lines upon retransfection of SW-13 cells with the recovered plasmids makes it difficult to establish causality between the expression of the hARF4p and bFGF secretion or clonal growth in soft agar. Nevertheless, these initialresults may indicate a link between the regulation of growth factor secretion and a member of the superfamily of regulatory GTP-binding proteins. Such a link is strengthened by the recent results in yeast which demonstrate a role for the ARF proteins in the protein secretory pathway and the immunolocalization of ARF to the Golgi cisternae in animal cells (12). Establishment of such a link, between two rapidly emerging fields of biological importance, should facilitate identification of the specific molecular details of ARF T. Serafini, R. A. Kahn, and J. Rothman, unpublished observations.
2613
action in protein secretion and of the regulation of bFGF secretion. Acknowledgements-We thank Andrea Cheville and Jose Acol for excellent technical assistance and Dr. Susan Flamm for assistance in constructing the ARF4 riboprobe vector. Dr. Tim Stearns provided original yeast strains and was an invaluable source of advice on all matters related to yeast.We thankDr. BillSugden for gifts of plasmids used in the constructionof the p266CH2 vector. Addendum-During completion of this manuscript it was noted that a recent publication (36) describes the partial cDNA sequence of a cloned open reading frame, called hARF4, obtained by hybridization with a human ARF pseudogene probe. The DNA sequence of this openreading frame is identicalto thatof human ARF4, described above. The derived protein sequences should therefore be identical; however, one error was noted in the published sequence at position 79 (Arg should be a Lys). REFERENCES 1. Kahn, R. A,, and Gilman, A. G. (1984) J. Biol. Chem. 259,6228-
6234 2. Kahn, R. A. (1991) Methods Enzymol. 195, 233-242 3. Schleifer, L. S., Kahn, R. A., Hanski, E., Northup, J. K., Sternweis, P. C., and Gilman, A. G. (1982) J . Biol. Chem. 257, 2023 4. Kahn, R. A. (1989) in G-Proteins (Birnbaumer, L., and Iyengar, R., eds) pp. 201-214, Academic Press, Orlando, FL 5. Kahn, R. A., and Gilman,A. G. (1986) J. Bid. Chem. 261,79067911 6. Kahn, R. A., Goddard, C., and Newkirk, M. (1988) J. Biol. Chem. 263,8282-8287 7. Stearns, T., Hoyt, M. A., Botstein, D., and Kahn, R.A. (1990) Mol. Cell. Biol. in press 8. Sewell, J. L., and Kahn, R. A. (1988) Proc. Nutl. Acud. Sci. U. S. A. 85,4620-4624 9.Price, S. R.,Nightingale, M., Tsai, S.-C., Williamson, K. C., Adamik, R., Chen, H.-C., Moss, J., and Vaughan, M. (1988) Proc. Nutl. Acad. Sci. U. S. A . 85, 5488-5491 10. Peng, Z., Calver, I., Clark, J., Helman, L., Kahn, R. A., and Kung, H. (1989) BioFactors 2 , 4 5 4 9 11. Bobak, D. A., Nightingale, M. S., Murtagh, J. J., Price, S. R., Moss, J., and Vaughan, M. (1989) Proc. Nutl. Acud. Sci. U. S. A. 86,6101-6105 12. Stearns, T.,Willingham, M.C., Botstein, D., and Kahn, R. A. (1990) Proc. Nutl. Acad. Sci. U. S. A. 87, 1238-1242 13. Okayama, H., and Berg, P. (1983) Mol. Cell. Biol. 3, 280-289 14. Sanger, F., and Coulson, A. R. (1975) J. Mol. Biol. 94, 441-448 15. Colbere-Garapin, F., Horodniceanu, F., Kourilsky, P., and Garapin, A.-C. (1981) J. Mol. Bid. 150, 1-14 16. Southern, P. J., and Berg, P. (1982) J. Mol. Appl. Genet. 1, 327341 17. Yates, J., Warren, N., Reisman, D., and Sugden, B. (1984) Proc. Natl. Acud. Sci. U. S. A. 81, 3806-3810 18. Yates, J., Warren, N., and Sugden, B. (1985) Nature 313, 812815 19. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979) Biochemistry 18, 5294-5299 20. Gubler, U., and Hoffman, B. J. (1983) Gene (Amst.) 25,263-269 21. Chen, C., and Okayama, H. (1987) Mol. Cell. Biol. 7,2745-2752 22. Melton, D. A., Krieg, P. A., Rehagliati, M. R., Maniatis, T.,Zinn, T., and Green, M. R. (1984) Nucleic Acids Res. 12, 7035-7056 23. Weiss, O., Holden, J., Rulka, C., and Kahn, R. A. (1989) J . Biol. Chem. 264, 21066-21072 24. Studier, F. W., and Moffatt, B. A. (1986) J . Mol. Biol. 189, 113130 25. Rosenberg, A. H., Lade, B. N., Chui, D., Lin, S.-W., Dunn, J. J., and Studier, F. W. (1987) Gene (Amst.) 56, 125-135 26. Johnston, M., and Davis, R. W. (1984) Mol. Cell. Biol. 4, 14401448 27. Sherman, F., Fink, G. R., and Lawrence, C. W. (eds) (1974) Methods in Yeast Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 28. Ito, H., Fukada, Y., Murata, K., and Kimura, A. (1983) J . Bucteriol. 153, 163-168 29. Bokoch, G. M., Katada, T., Northup, J. K., Hewlett, E. L., and Gilman, A. G. (1983) J. Biol. Chem. 258, 2072-2075
2614
ADP-Ribosylation Human
30. Schaffner, W., and Weissman, C. (1973) Anal. Eiochem. 56,502514 31. Leibovitz, A., McCombs, W. B. I. I. I., Johnston, D., McCoy, C. E., and Stinson, J. C. (1973) J. Natl. Cancer Inst. 51, 691-697 32. Wellstein, A., Lupu, R., Zugmaier, G., Flamm, S. L., Cheville, A. L., Delli Bovi, P., Basilico, C., Lippman, M. E., and Kern, F. G . (1990) Cell Growth 1, 63-71 33. Towler, D. A., Eubanks, S. R., Towery, D. S., Adams, S. P., and
Factors Glaser, L. (1987) J. Biol. Chern. 262, 1030-1036 34. Gunning, P., Mohun, T., Ng, S., Ponte, P., and Kedes, L. (1984) J. Mol. Evol. 20,202-214 35. Farnham, P. J., and Means, A. L. (1990) Mol. Cell. Biol. 10, 1390-1398 36. Monaco, L., Murtagh, J. J., Newman, K. B., Tsai, S., Moss, J., and Vaughan, M. (1990) Proc. Natl. Acad. Sci. U. S. A . 87, 2206-2210
