Characterization of two cDNAs encoding folate-binding proteins from ...

Vol. 266, No. 26. Issue of September 15, pp. 17243-17249,1991 Printed in 1J.S.A.

THEJOURNALOF BIOI.OGICAL CHEMISTRY

‘0 1991 by The American

Society for Biochemistry and Molecular Biology, Inc.

Characterization of Two cDNAs Encoding Folate-binding Proteins from L1210 Murine Leukemia Cells INCREASEDEXPRESSIONASSOCIATEDWITH

A GENOMICREARRANGEMENT* (Received for publication, January 25, 1991)

Kevin E. Brigle, EricH. Westin, Micah T. Houghton, and I. David Goldman$ From the Departments of Medicine and Pharmacology, Virginia Commonwealth University, Medical College of Virginia, Richmond, Virginia 23298

L1210 murine leukemic cells grown under condi- properties of bidirectional uphill transport (2-4) and sensitivtions of continuous low folate concentrations acquire ity to structurally unrelatedinorganic and organic anions (5increased levelsof a high affinityllow capacity folate- 7). While the influxof folic acid is mediated in small part by binding protein (FBP). Using an oligonucleotide probe the tetrahydrofolatecofactor carrier, its transportoccurs precomplementary tothe human FBP, we have cloned and dominantly by mechanism(s) which aredistinct from this sequencedtwomurine FBP cDNAs isolatedfrom a route (8-11). Sirotnak et al. (8, 9) have demonstrated the library constructed using a L1210 subline adapted for presence of an anion-insensitive low affinitylhigh capacity growth on low levelsof 5-formyltetrahydrofolate. The transport system which represents the primary routeof folic encoding proteins, designated FBPl and FBP2, have acid influx at micromolar concentrations in L1210 cells. Henpredicted M , values of 29,415 and 28,821, respec- derson and Strauss (12) have recently described an additional tively. These proteinsshare significant sequence identity with each other (70%)and with the deduced se- transport system operating in murineleukemic cells which is quences of the human- andbovine-encoded FBPs (68- manifested at low pH and mediates the energy-dependent 79%).Southern blot analysis of the low folate-adapted influx of folate, 5-formyltetrahydrofolate, and methotrexate cell line revealed that, while neither of the two FBP- in the 5-10 PM range. In addition to these transport routes encoding genes was amplified, Fthe B P l genomic locus which are characterized by rapid association and dissociation hadundergonerearrangement. On Northern blot of substratewith arapidly translocatingcarrierelement, analysis, this rearrangement was reflected in the en- L1210 sublines (13,14)have been identifiedthrough selection hancedexpression(>lOO-fold) of a FBP1-specific a t low folate levels that acquire their folates via the reversible, transcript which was 200 nucleotides shorter than the increased expressionof a slowly cycling membrane-associated corresponding L1210parentaltranscript.Theinhigh affinity folate-binding protein. These proteins have been creased expression of this transcript coincided with isolated from a variety of sources (reviewed in Ref. 15) and the increased expression of a membrane protein (Mr = have beenshown to be anchoredtothemembrane by a 38,000) which could be affinity-labeled with a N-hy- glycosyl-phosphatidylinositol linkage (16, 17). Recent studies droxysuccinimide ester of [3H]folate. Accordingly,the (18,19)using MA104 cells have demonstrated that the folateFBPl transcript appears to encode the high affinity/ binding protein recycles during folate utilization, traversing low capacity FBP. Compared to parental L1210 cells, the cell viaendocytosisina novel uncoatedpitpathway. expression of the FBP2-encoding transcript was un- Screening of human cDNA libraries from several different changed in this cell line and, while the exact nature of tissues identified two distinct yet closely related folate-bindthe protein is unclear, FBP2 may represent a fetal form ing proteins, one of which appears to represent a fetal form of the FBP. of the protein (16, 20-22). While each of the transport mechanisms described above has beenidentified and characterized in murine leukemic cells Membrane transport of folate compoundsin L1210 murine primarily on the basisof their individual kinetic parameters, which comprise these systems leukemia cells has been the subject of intensive studieswhich none of the component proteins have provided evidence for the existence of several mecha- in L1210 cells has been isolated or directly characterized ata nisms of (anti)folate uptake. The transport of methotrexate molecular level. To this end,we have begun an effort toclone and other antifolates plays a critical role in drug selectivity and characterize thosegenes that encode the various proteins and cytotoxicity and, a t pharmacological levels, is mediated which either comprise or are associated with these transport by a high affinity/low capacity tetrahydrofolate cofactor car- mechanisms. In this paper,we describe the cloning, sequence determination, and characterization of two distinct folateriersystem (1) which displays theuniquethermodynamic binding protein-encoding cDNAs from a library constructed using a L1210 subline which was adapted for growth on low * This work was supported by National Cancer Institute Grants CA-39807 and CA-09340. The costs of publication of this articlewere levels of 5-formyltetrahydrofolate. Additionally, we describe defrayed in part by the payment of page charges. This article must a potential mechanismfor increasing expression of one of the therefore be hereby marked “aduertisement” in accordance with 18 encoded folate-binding proteins via a rearrangement of the U.S.C. Section 1734 solely t o indicate this fact. encoding genomic locus. The nucleotide sequence(s)reported in this paper ha$been submitted totheGenBankTM/EMRLDataBankwith accession number(s) M64782 and M64817. $ T o whom correspondence and reprint requests should be addressed: Dept. of Medicine,Medical College of Virginia, Box 230 MCV Station, Richmond, VA 23298. Tel.: 804-786-9723.

MATERIALSANDMETHODS

Chemicals-[3’,5’,7, 9-:’H]Folic acid was obtained from Moravek Biochemicals and purified by high performance liquid chromatography (23) prior to use. Restriction enzymes were obtained from Be-

17243

17244

Characterization of Two Murine Folate-binding Protein cDNAs

thesda Research Laboratories (BRL) or New England Biolabs. All of the indicated restriction enzyme/gg of DNA. Ten-microgram aliother reagents were obtained in the highest purity available from quots were fractionated on0.8% agarose gels, denatured, neutralized, various commercial sources. and blotted onto the Nytran membrane (31). Hybridization, washes, Tissue Culture-Murine L1210 leukemia cells were grown in RPMI and autoradiographywere done as described for Northern analysis. 1640 medium supplemented with 10% dialyzedbovine serum (HyI n Vitro Transcription and Translation-cDNA inserts were subclone), 2 mM glutamine, 20 p M 2-mercaptoethanol, penicillin (100 cloned into the appropriate pBluescript vectors such that the assigned units/ml), streptomycin (100 pg/ml), and 2.2 p~ folic acid. Cell line ATG start codon for each predicted polypeptide would be the 1st LL1 was developed by firstdepleting L1210 cells of intracellular methionine residue encoded on a transcript initiated from one of the folates by culturing for 10 days in folate-free RPMI 1640 medium two promoters (T3 or T7) present in thesevectors. According to the supplemented asabove with the substitutionof GAT (200 p~ glycine, manufacturer's (Stratagene) instructions,1.0 pg of linearized plasmid 100 p M adenosine, and 10 p~ thymidine)in place of folic acid. was used as the template to synthesize a m7GpppG-capped transcript Following removal of GAT, 1 nM 5-formyltetrahydrofolate was added encoding the predicted polypeptide. Aliquots of each reaction were as the sole folate source. In the initial month, cells grew very slowly fractionated bygel electrophoresis on 1.0% formaldehyde-agarose and showed irregular morphology. Following this period, cells were gels in order to determine the size, integrity, and concentration of passaged twice weekly, and the folate concentration was decreased the transcripts. The resultant RNA (2.0-3.0 pg) was translated in stepwise over the following 2 months to a final maintenance level of vitro using a rabbit reticulocyte translation system (Promega) con0.4 nM. At this point, when the developing cell line (LL1) had a taining ["S]cysteine in the presenceor absence of canine pancreatic growth rate approximately40% of the parental L1210 line, cells were microsomal membranes (Promega), and the translation products were harvested for construction of the cDNA library. This line was not analyzed ona 12.5% reducing SDS-PAGE gel (32). The gel was cloned prior to construction of the cDNA library or bindingstudies. processed with EN'HANCE (Du Pont-New England Nuclear) accordAll cells were cultured for a t least six generationsin folate-free ing to manufacturers instructions, dried, and autoradiographedfor 12 medium prior to binding studies. h to visualize the translation products. Construction and Probingof cDNA Library-Cells (2.5 X 10') were Affinity Labeling and Glycanase Treatment of Cell Membranes-Nharvested, washed with ice-cold saline, and immediately solubilized Hydroxysuccinimide (NHS)-[3H]folic acidwas prepared by a modiin GMST (4.0 M guanidine thiocyanate, 7% 2-mercaptoethanol, 2% fication of the method of Henderson and Zevely (33) as previously N-lauroylsarcosine, 10 mM Tris, pH 7.4) essentially as described by described (34). For affinity-labeling studies, 2 X lo8 cells were harwas prepared by CsCl density vested, washed with saline, and resuspended in HBS mM Cbirgwin et al. (24).TotalRNA (20HEPES, centrifugation, and poly(A)' RNA was selected by chromatography 140 mM NaCl, 5 mM KCI, 2 mM MgCL (pH 7.4)). NHS-["Hlfolic acid (25) using oligo(dT)-cellulose (BRL). Blunt-ended, double-stranded was added to a concentration of 200 nM in a final volume of 2.0 ml, cDNA was prepared from 7.5 pg of poly(A)+ RNA using the RNase and the cells were incubated for 10 min at 0 "C. The cells were H method(26) according totheBRLcDNASynthesisSystem's recovered by centrifugation, washed with HBS, and resuspended in instruction manual and ligated with BstXI linkers (Invitrogen). The 4.0 ml of carbonate lysis buffer (1 mM NaHCOS, 2 mM CaC12,5 mM linkered cDNA was size-fractionated by gel electrophoresis, and the MgC1, (pH 7.8)) with 2 mM phenylmethylsulfonyl fluoride a t 0 "C. fraction above 500 bp was isolated by electroelution witha Schleicher Cells were then disrupted in a Dounce homogenizer, layered onto a and Schuell Elutrap. The resultant cDNA was inserted into BstXI- discontinuous sucrose gradient(2.0 ml of 70% sucrose/6.0 ml of 42% cut pCDNAII (Invitrogen)andtransformedinto Escherichia coli sucrose), andcentrifuged a t 20,500 rpm for 1 h in a SW41 rotor. The strain Epicurian (Stratagene) producing a library of approximately opaque band at the homogenate/42% sucrose interface was removed, 7.3 X lo5 primary recombinant clones. Colony lifts (Schleicher and diluted to 10ml with carbonate buffer, and centrifuged a t 30,000 X g Schuell) were prehybridized a t 37 "C in 5 X SSC, 50 mM NaPO, (pH for 30 min. The pellet was resuspended in 200 p1 of 0.5% SDS, 100 6.4), 0.1% SDS,' 5 X Denhardt's solution, 250 pg/ml total yeast RNA mM 2-mercaptoethanol and adjusted to2 pg/pl protein. Fifty-micro(Boehringer Mannheim), and 40% formamide, followed by hybridi- gram aliquotswere boiled for 3 min, resuspended in 50 pl of glycanase zation with a"'P-end-labeled oligonucleotide (1X lo6cpm/ml). Filters assay buffer (0.2 M sodium phosphate, 10 mM 1,lO-phenanthroline were washed five times with 2 X SSC, 0.2% SDS at 50 "C prior to hydrate, 1.25% Nonidet P-40 (pH 8.6)) in the presence or absence of autoradiography a t -80 "C. Oligonucleotides used were as follows: 5'0.5 unit of N-glycanase (Genzyme), and incubatedfor 18 h at 37 "C. CTCGTAGAGGCAGGTGTCCTGGATGAAATGCCGTTTGCA-3' The labeled membrane preparations were fractionated on a 12.5% and 5'-CCAGTTCCAGCCCTTGTGCCAGTTGCTCTTGCAGGTreducing SDS-PAGE gel, cut into 2-mm slices, and dissolved overGTA-3'. night in Soluene 350 (Packard). Hionic (Packard) scintillation fluor Subcloning and Sequence Analysis-Inserts from the cDNAclones chosen for furtheranalysis were isolatedand subcloned into wasadded to the samples, and radioactivity was determined in a liquid scintillation spectrometer (Beckman Instruments). pBluescript vectors (Stratagene). Double-stranded DNA sequencing (27) was performed by the dideoxy chain termination method (28) RESULTS using the Sequenase2.0 sequencing kit from United States Biochemicals Corp. Using either internal restriction sites or synthetic 17-mer Strategy and cDNA Cloning-A cDNAlibrary was conprimers, theclones were sequenced in both directions with the excep- structed from poly(A)+ RNA isolated from cell line LLI, a tion of the first 26 bp of pKD18 and the first 81 bp of pKD21. The nucleotide sequencesin theseregions were unambiguously determined L1210 subline which had been growing for several months on a reduced concentration of 5-formyltetrahydrofolate. Similar by three independent reactions, Computer analyses of the nucleotide selection schemes by other researchers have resulted in cell and protein sequences were performed using the sequence analysis software package of the University of Wisconsin Genetics Computer lines which overproduce either a folate-binding proteinor the Group (29). reduced folate transport system (13, 14, 35). In preliminary Northern Blots-For blot analysis, 3.0 pg of poly(A)' RNA was experiments, two 39-mer oligonucleotide probes complemenfractionated by gel electrophoresison 1.0%formaldehyde-agarose gels. The gels were stained with ethidium bromide to ascertain even- tary to several regions of the human folate-binding protein ness of loading, transferred onto Nytran membrane (Schleicher and (16, 22) were synthesized and, in Northern analyses, hybridSchuell), andfixed by baking in a vacuum oven at 80 "C for 2 h (30). ized to several distinct transcripts (0.5-8.0 kilobases) present Hybridizations were done as described above except that the hybrid- in poly(A)+ RNA isolated from cell line LL1. Therefore, 2.5 ization solution was 50% formamide a t a temperature of 42 "C. For X IO5 primaryrecombinants were screened under similar nick-translated probes (1 X lo6 cpm/ml), 10% dextran sulfate was hybridization conditions, and 34 potential positive clones of added to the hybridization mixture. The filters were washed three times with 0.1 X SSC, 0.1% SDS at 60 "C, and autoradiographywas various signal intensities were identified. Following restriction analyses and additional probing, these clones fell into from 6 h to 6 days a t -80 "C with an intensifyingscreen. Southern Blots-Genomic DNA was digested for 18 h with 3 units two distinct groups. The largest groupcomprised 31 clones all

of which exhibitedstrong

hybridizationsignals.

For both

The abbreviations used are: SDS,sodium dodecyl sulfate; bp, base groups, the clone containing the largest insert (pKD18 and pair(s);PAGE, polyacrylamide gel electrophoresis; HEPES,4-(2hydroxyethy1)-1-piperazineethanesulfonic acid; NHS, N-hydroxysuc- pKD21, respectively) was chosen for sequence analysis. cinimide; CHAPS, 3-[(3-cholamidopropyl)dimethylammonio]-l-pro- Nucleotide and Amino Acid Sequence-Both strands of the cDNA inserts present in pKD18 and pKD21 were sequenced panesulfonic acid; nt, nucleotide(s).

Characterization of Two Murine Folate-binding ProteincDNAs employing the sequencing strategies summarized in Fig. 1. The complete nucleotide sequence and predicted amino acid sequence for both inserts is given in Fig. 2. The 979-bp pKD18 insertcontains a765-bp openreadingframe(designated FBP1) which is flankedby 59- and 155-nucleotide 5 ’ - and 3’noncoding regions, respectively.The 3“noncoding region contains a polyadenylation signal located 34 bp upstream from the poly(A) tail. The ATG start codon was assigned according to its similarity to the eukaryotic consensus start sequence as described by Kozak (36). While the 5”noncoding sequence contains several termination codonsin the two closed reading frames preceding this assigned ATG start, there are no preceding in-frame termination codons, and it is possible that thepredictedreadingframeactuallyinitiatesfurtherupstream. The 1223-bp pKD21 insert encodesa753-bpopen reading frame (designated FBP2)which is flanked by 5’- and 3”noncoding regions of 272 and 198 nucleotides, respectively. The translation initiation site isassigned to the ATG codon starting atnucleotide 273 as this is the first methionine codon downstream from four in-frame termination codons. In addition, thenucleotide sequencesurrounding thiscodon is similar to the eukaryotic consensus start sequence, and it is identical with the sequence surrounding the ATG start codon assigned for FBP1. The 3“noncoding region contains a polyadenylation signal located 34 bp upstream from the poly(A) tail. The two cDNA clones show only very limited sequence identity within their5‘- and 3’-untranslatedregions. Codon usage tables computed for each protein (not shown)reveal distinct protein-specific differences in the preferredcodons which are used to specify arginine, glutamine, glycine, and threonine residues. FBPl and FBP2 each possess typical peptide signal sequences and have predicted molecular weights (Mr = 29,415 and 28,821, respectively) similar to those of other sequenced folate-binding proteins (16, 20-22, 37). As a murine-encoded folate-binding protein has not yet been purified, the molecular weights and amino terminiof the processed, mature forms of FBPl and FBP2 are unknown. Based on the deduced sequences of corresponding folate-binding proteins from human cell lines, the leader peptides of FBPl and FBP2 may range in length from 16-25 residues. The putative signal sequences of these two proteins are dissimilar in residue composition andpredictedcharge whichmaysuggest targeting of the proteinstodifferent cellular compartmentsormembrane types. As shown in Table I and in Fig. 3, FBPl and FBPS have significant sequence identity with each other and with the deducedsequences of the human- and bovine-encoded folate-bindingproteins.Interestingly, FBPB sharesgreater identity with the human placental folate-binding protein than with FBP1. Likewise, F B P l shares greatest identity with the nonplacental proteins. In a similar manner, both FBPS and S

E

S

8 8

pKDl8

17245

-141

- 76 -11 54 119 184 249 314 379 444

509 5 74 639

- FBPl

A .

704

~

~

-

769

-

~

834 8W

wa t9

r i

B

E

0

x

pKD21

-

F8P2

~

FIG. 1. Restriction map and sequencingstrategy for cDNAs encoding the folate binding proteins FBPl (pKD18) and FBPB (pKD21). B = BamHI, E = EcoRI, S = SstI, X = XbaI.

FIG.2. cDNA clones encoding themurine folate-binding proteins. A , FBP1-encodingpKD18cDNA clone; B , FBP2-encoding pKD21 cDNA clone. For each clone, the nucleotide sequence is numbered on the left and the protein sequence is numbered on the right. The A of the initiating methionine codons are designated as nucleotide number 1. Termination codons are indicated by an asterisk and the polyadenylation signals are underscored.

17246

Characterization of

Two Murine

TABLE I Percent identity calculated the for deduced amino acid sequences of FHI'I and FHP2 versus the deduced sequences for the human-(16, 20. 22), human placental- (21), and bovine- (37) encoded folatebindina wrotein FRPl

FRPl FRP2

FBPl FEPZ HUMAN PLACEN BOVINE

FBPZ

Human

Human Placental

Bovine

70

79 70

72 77

73 68

70

1 mahlrntvqllllwnmsec..aqs.ratraRTELLNVCWd rdRTDLLNVCWd maukqtplII.lvymvttgs g m a q r r n t t q l l l l l v u v a w ge a q t . r i a u a R T E L L N V C M n mvukumplllllvcvatms aqdRTDLLNVCMd .aqaprtpraRTDLLNVCUd

....................

....... .......

57 AKHHKekPCPEDnLHDOCsP AKHHKtkPGPEDkLHDOCSP AKHHKekPGPEDkLHEOCrP AKHHKtkPGPEDkLHDOCSP AKHHKaePGPEDsLHEOCsP

58 117 FBPl UktNsCCstmqEahKdiSyLYrFNYnHCCtMtseCKrH FlPDtCLYECSPWLGPYIqq FBPZ UkkNaCCsv@qElhKadS rLY.FWYdHCGkMepaCKsH FlODsCLYECSPWLCPYIqq HUMAN U r k N a C C s t ~ q E a h K d v SyLYrFNUnHCGeMapaCKrHF10DtCLYECSPWLGPUlqq PLACEN UkkNaCCtaSTSqElhKdtSrLYnFNUdHCGkUepaCKrHFlODtCLYECSPNLGPUlqq BOVINE U r k N a C C s v a i E a x K d i S yLYrFNUdHCGkUepaCKrHFlODtCLYECSPNLGPUlre

FBPl FBP2 HUMAN PLACEN BOVINE

177 118 VdQsURKERiLdVPLCKEDC qqVVEdCqsSfTCKsnUHkG Un&SGhNeCpvgAschpFt VdQsURKERfLdVPLCKEDC hqVVEaCrtSfTCKrcUHkGUdYsSGiNkCpntApchtFe VdOsURKERvLnVPLCKEDC eqVVEdCrtSyTCKsnYHkCUnVfSGfNkCavgAacqpFh Vn(rfURKERfLdYPLCKEDC qrWEdChtShTCKsnYHrGUdVtSGvNkCpagAlcrtFe VnQrURKERvLgVPLCKEDC qsWEdCrtSyTCKsnUHkG UMJSGyNqCpvkAahcrFd

178 236 FEP1 fYFPTsaaLCeeiYsHSYKlS~RGSGRClOMFDpaOG NPWEEVarFYAearn...sGa FBP2 yYFPTpasLCeglUsHSYKvS~RGSGRCIOMUFDstOGNPNEDVvkFYASfmt..sCt HUMAN fYFPTptvLCneiYtHSYKvS~RCSGRClOMYFDpaOGNPNEEVarFYAaam...sGa PLACEN SYFPTpaaLCeglUsHSYKv S~RGSGRClOnUFDsaOCNPNEEVarFYAaamhvnaGe BOVINE fYFPTpaaLCneiUsHSYKvS~RGSGRCIOMUFDpfOGNPNEEVarFYAefi..sGs

235

FBPl FBP2 HUMAN PLACEN BOVINE

255 g l h g t u p l l c s l s l v l . l u vi s v p h a a v l l v p s l a p v l s l u l pg gpwaaupfllslalrnl.lul I s mlhgtggltlslalmlqlul l g tpqgi

..................

FIG. 3. Alignment of F B P l and FBP2 with otherfolatebinding proteins. Amino acid residues conservedin all five proteins are shown in uppercase. Dots representalignmentadjustmentsto provide best fit. Potential N-glycosylation sites are underscored, and the conserved glycosylation site is indicated bv an asterisk.

Folate-bindingProtein cDNAs

--

1 2 3 4 5 6 3 2 3 4 5 6

12.2. 8.2. 6.1-

"

.

4.Q-

2.0.

-.

..

.. "".Y

, .

"

"

B

A

FIG. 4. Southern blot analysis of L1210 and LLl genomic DNA. L1210 (lanes 1.3, and 5 ) and LL1 (lanes 2 , 4 , and 6 ) genomic DNA was digested with HindIII (lanes I and 2), XbaI (lanes 3 and 4 ) , and RQ&I (lanes 5 and 6 ) . Panel A, hybridization with radiolabeled SstI fragment from pKD18 (FBP1 probe). Panel R, hybridization with radiolabeled RamHI fragment from pKD21 (FRP2 probe).

4.91.9

1.3-

* I,

FIG. 5. Northern analysis of L1210 and LL1 poly(A)+RNA. RNA extracted from LL1 (lanes I and 3) and L E 1 0 (lane 2) was hybridized with the radiolabeled Sstl fragment from pKD18 (FRP1 probe). Lane I , 6-h exposure of LL1; lane 2,&day exposure of L1210; lane 3,6-day exposure of lane 1. The blot was stripped, reprobed with the radiolabeled HamHI fragment from pKD21 (FRP2 probe), and exposed for 6 days: lane 4 , L1210, lane .5, LLl.

suggesting that a genomicrearrangement hasoccurred inLL1. In support of this notion, one of the three bands detectedin the XbaI digests and one of the two bands detected in the BarnHI genomic digests of these cell lines are of different sizes (Fig. 4A, lanes 3-6). Based on the signal intensity of the the placental protein contain two potential N-linked glyco- detected bands, the pKD18 hybridizing sequences appear to sylation siteswhile the other proteins each contain three such be present in both genomes inequal copy number. The sites. Only one of these sites (correspondingto FBPl residue presence of three hybridizing bands in the XbaI digests (no Asn-202 in Fig. 3) is conserved among all five proteins. There cleavage sites in the encoding cDNA or probe) is consistent are 18 conserved cysteine residues and 10 conserved trypto- with either the gene having several introns or with the existphan residues among these proteins. Theconserved cysteine ence of several closely related genes. These blots were then residues have been suggested to play a structural role within stripped and probed for the FBP2-encoding gene using the the protein via the formation of disulfide bonds, while the pKD21 BamHI fragment (Fig. 4 R ) . In addition to detecting tryptophan residues have been proposed to function as poten- several unique bands in each genomic digest, this probe also tial ligand binding sites(38).Computeranalyses utilizing cross-hybridized with one previously detected pKD18-hybridChou and Fasman (39) criteria predict very similar secondary izing band in each lane of the blot (denoted by bullets in Fig. 4A, lunes 1-6). This cross-reactivity is likely due to the strong structures for each of the five folate-binding proteins (data not shown).Also, excluding the putative signal peptides, Kyteprimary sequence identity observed in the amino-terminal and Doolittle (40) analyses predict each protein tohave only coding region of the two cDNAs. Disregarding the crossa single region of extended hydrophobicitylocated at the hybridizing signals, the bands detectedby the pKD21 probe carboxyl terminus that appears capable of interacting with in the genomic digests of both cell lines were of equal size and the cell membrane. This domain hasbeen suggested to act as signal intensity. As with the FBP1-encoding gene, the presence of more than one hybridizing band in the RamHI and the recognition site for anchoring the protein to the membrane XbaI genomic digestsis consistentwith either thegene having via a glycosylphosphatidylinositol linkage (16, 17). Southern Analysis"L1210 andLL1 genomic DNA was several introns orwith the existenceof several closely related digested with restriction enzymes HindIII, XbaI, and BamHI genes. Further, upon a 6-day exposure, both cDNA probes and used in Southern analysis to determine the complexity of detected several additional unique bandsof weak signal intenthe two genes encoding FBPl and FBP2. The radiolabeled sity (not shown). Northern Analysis-The probes used for Northern hybrid500-bp SstI fragment of pKD18 (FBP1) and 600-bp RarnHI fragment of pKD21 (FBP2) were used as probes (see Fig. 1). izations were the same as those used in Southern analyses. Neither of these probes was gene-specific, and, even under As shown in Fig. 5, probing of LL1-derived poly(A)+ RNA the conditions of high stringency usedin this experiment, with thepKD18 SstI (FRP1)fragment revealed asingle, some cross-hybridization occurred. As shown in Fig. 4A, lanes strongly hybridizing 1200-nt transcript which was evident 1 and 2, the pKD18 SstI (FBP1) probe hybridized to a single after only 6 h of exposure (lane 2), while two additional ,band in both genomic DNAs cleaved with HindIII. However, transcripts (2500 nt and 7200 nt) became evident after 72 h the hybridizing band in LL1 was 1800 bp smaller than the of exposure (shown in a6-day exposure in lane3 ) .In contrast, this probe to identify a band detectedin L1210 (7700 bp uersus 9500 bp, respectively), a 6-day exposure was required for

Characterization of Two Murine Folate-binding Protein cDNAs

17247

single 1500-nt transcript (lane 2 ) in poly(A)+ RNA isolated from parental L1210 cells and neither of the larger mRNA species could be detected. The prominent 1200-nt LL1 trans43cript was expressedseveral hundredfoldgreaterthanthe 29longer, 1500-nt parental transcript. It is possible that the shorter length and enhanced expression of this prominent LL1 transcript may in some way be causally related to the 18genomic rearrangements observed in the Southern analyses shown in Fig. 4A. When the Northern blots were stripped and probed with the pKD21 RarnHI (FBP2) fragment, each cell FIG. 6. In vitro translation and processingof thetwo line revealed a diffuse 1500-2000-nt band (which may poten- cDNA-encoded folate-binding proteins. RNAs transcribed from tially resolve into two distinct bands) visible after 6 days of pHhescript vectors containing the cDNA inserts (described under were translated in rabbit reticulocyte exposure (Fig. 5, lanes 4 and 5 ) . Unlike the pKD18 comple- "MaterialsandMethods") extracts in the presence of ["Hlcysteine. The translation products mentary transcripts, the presence of this diffuse band was not were resolved by SDS-PAGE and detected by fluorography. Lane 1, increased in the LL1line. In fact, theL1210 line containeda control reaction which contains no RNA; lanes 2 and 3, products of slightly higher level of this transcript than the LL1 line. A the FBP1-encoding vector translated with and without microsomal membranes, respectively; lanes 4 and -5,products of the FBPB-encoddigital scan of the autoradiograph resulting from hybridization of this blot with an actin probe revealed that the LL1 ing vector translated with and without microsomal membranes, remRNA was present a t a ratio of approximately 1.5:l uerscu spectively. the L1210 mRNA. This minor difference in loading cannot account for the greatly increased abundance observed for the ,18.4 14.3 29.0 43.0 m.0 97.4 I1 I I I I I I F B P l specific transcript. While the FBP1- and FBP2-encod20.0h I ing transcripts each appear to be present in parental L1210 cells in relatively equal amounts, they are not present in any appreciable quantity. In thelow folate-adapted LL1line, only the FBP1-specific transcript is present a t increased levels (alongwithtwolargercross-hybridizing mRNA species of unknown identity). Based on the observed size of the mRNA species detected in the Northern blots, it is apparent that ;