- Email: [email protected]
Multiple folate transport systems in L1210 cells
Advan. Enzyme Regul., Vol. 32, pp. 3---15,1992 Printed in Great Britain. All rights reserved
0065-2571/92/$15.00 © 1992 Pergamon Pressplc.
MULTIPLE FOLATETRANSPORT SYSTEMS IN L1210CELLS J. FAN, K. S. VITOLS and F. M. HUENNEKENS Division of Biochemistry, Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, CA 92037
INTRODUCTION: TRANSPORT AS A CRITICAL PARAMETER THE CHEMOTHERAPEUTIC EFFICACY OF METHOTREXATE
IN
Methotrexate (MTX)~, the 4-amino-10-methyl analog of folic acid, was one of the first agents shown to be effective in cancer chemotherapy, and it continues to be used extensively for the treatment of choriocarcinoma, acute lymphocytic leukemia, and, in combination with other agents, for maintenance of drug-induced remissions. Extensive information is available for MTX with regard to its transport into cells, metabolic conversion to polyglutamate forms, and inhibition of dihydrofolate reductase (DHFR; EC 1.5.1.3). Resistance to MTX, the result of defective transport or polyglutamylation, of increased levels of the target enzyme DHFR, or of mutations in DHFR, has been documented in cellular model systems and in patients undergoing drug therapy. In addition to its intrinsic value, this large body of information about MTX has provided valuable guidance for parallel studies with other drugs. Recognition of transport as a critical parameter in the chemotherapeutic efficacy of MTX began with the classic report of Kessel, Hall and Roberts in 1965 (1). These investigators, using a series of murine tumor cell lines, found that a linear relationship existed between the rate of uptake of M T X in vitro and the ability of the drug to kill these cells in vivo. Subsequent studies in a number of laboratories, conducted primarily with L1210 murine leukemia cells, demonstrated that MTX was taken up by an active transport system whose normal substrates were folate compounds, and the substrate specificity and kinetic characteristics of folate transport systems in various cells and tissues were delineated. This work, which utilized intact cells, was able to provide only limited information about components, mechanism, or regulation of the transport systems. In 1977, however, a membrane-associated folate binding protein, shown to be essential for folate transport, was purified to homogeneity from Lactobacillus casei (2) and, more recently, folate transporters from KB cells (3), MA-104 cells (4) and human placenta (5) have been isolated and characterized.
4
J. FAN, et al. G E N E R A L P R O P E R T I E S OF F O L A T E T R A N S P O R T SYSTEMS OF L1210 CELLS
An unexpected complication in folate transport has emerged from the evidence that certain eukaryotic cells can express two different systems for this purpose. L1210 cells, for example, when propagated in vitro on the usual laboratory media containing m i c r o m o l a r concentrations of folate, express a system whose primary substrate is 5-methyltetrahydrofolate (K t ~ 1 /~M). 5-Formyltetrahydrofolate (folinate) and M T X are also good substrates for this system, as indicated by their K t values of ca. 5/~M, but folate (K t > 100/~M) is a poor substrate (reviewed in (6)). Alternatively, growth of L1210 cells on nanomolar concentrations of folate or folinate leads to the gradual expression of a separate system (7) in which folate is the primary substrate (K t < 1 nM). 5-Methyltetrahydrofolate, 5-formyltetrahydrofolate and M T X are transported less efficiently, although their K t values are still in the nanomolar range. Based upon their kinetic characteristics, these will be referred to hereafter as the L1210 "nM" and "/~M" folate transport systems. Other investigators have preferred the term "reduced folate/MTX" for the latter system. Still others have designated the two systems as "high affinity/low capacity" (nM) and "low affinity/high capacity" (/~M). Properties of these folate transport systems are summarized in Table 1. Although the /~M transporter is expressed in a much smaller amount (ca. 60-fold) than the nM transport protein, the higher rate of turnover (ca. 7000-fold) results in a 100-fold higher Vmax. Thus, the overall efficiency of the ~M
TABLE 1. FOLATE TRANSPORT SYSTEMS IN L1210CELLS Property
~M System
nM System
Kt Vmax,pmol/min/106cells Turnover number*
mFH4 > fFH4 -- MTX > F 1 ~tM 5 ttM 100/zM 0.5 5
F > mFH4 = FH4 > MTX <1 nM 10 nM 100 nM 5 x 10-3 7 x 10-4
Amount, pmol/106 cells number/cell Molecular weight, kDA
0. l 6 x 104 43
6 3.6 x 106 39 Yes Yes
Substrate specificity
CHOt GP.t Energy source Mechanism Regulation
No No Anion gradient Anion exchange cAMP
? Endocytosis ? Tandem-/tM transporter ? ?
*Molecules/min/transporter. tAsparagine-linked carbohydrate detected by treatment with peptide:N-glycosidaseF. ~Glycosylphosphatidylinositol.
FOLATE TRANSPORTSYSTEMSIN LI210CELLS transport system is much greater than that of its nM counterpart. A similar coexistence of the nM and/~M folate transport systems has been observed in CCRF-CEM human leukemia cells (8), and the two transport systems have been characterized individually in other eukaryotic cells. Prior to the present study, relatively little was known about the molecular properties of the L1210/£M and nM folate transporters, although labeling experiments had indicated that the molecular weight of the former was 43--45 kDa (9-11). This paucity of information was largely due to the low level of the/~M transporter and to the difficulty in obtaining sublines upregulated for expression of either transporter. The present investigation was undertaken, therefore, to develop a rapid and efficient procedure for isolating these proteins. Biotin derivatives of MTX and folate containing a dissociable S-S linker have been synthesized and used to label the transporters in intact cells (12). Streptavidin-agarose beads are employed to collect the labeled proteins from detergent extracts of the isolated cell membranes, and the proteins are released by reduction of the S--S bond in the probes. SYNTHESIS O F B I O T I N D E R I V A T I V E S FOLATE
OFMETHOTREXATE
AND
The prototype compound in this series was biotin-SS-MTX, whose structure is shown in Figure 1. The structure is similar to fluorescein-MTX (9, 13), except that the spacer group is lengthened by the addition of a dissociable disulfide bond and fluorescein is replaced by biotin. The synthetic procedure for biotin-SS-MTX is outlined in Figure 2. MTX was converted to the N-hydroxysuccinimide (NHS)-ester, and the latter was reacted with diaminopentane (DAP). The product MTX-DAP, after purification by ion-exchange chromatography, was linked to the commercially available N-hydroxysulfosuccinimide (NHSS)-ester of a biotin compound containing a disulfide group. This sequence was modified from that used for the preparation of F-MTX (viz., FITC + DAP ---> FDAP ---> F-MTX) because of the limited availability of NHSS-SS-biotin and the non-chromophoric nature of biotin. Biotin-SS-MTX was obtained in an overall yield of ca. 50%, and since only 50-60% of the compound
","~" ,y% NHI
~
coo. !
,~, ~-"-(¢"')'-"- .c. g-Cc.,),-s-s -(c.,),-..- c-Co.,).
FIG. I. Structure of biotin-SS-MTX.
J. FAN, et al.
MTX 0 x) (a) EDC, lh, r.t. (1.2x) (b) NHS, 4h, r.t. (1.2x)
DAP
(lOx)
+
MTX-NHS Ox)
l
(a) 3h, r.t. (b) ppt. w/acetone (c) DEAE-TrlIacryl (H,O)
MTX-DAP
+
NHSS-SS-Biotin
(Ix)
(lx)
I (a) 6h, r.t, I (b) ppt. w/acetone ] (c) DEAE-TrlIecryl ~L (0.05M NH4HCO,) m
Biotin-SS-MTX FIG. 2. Flow diagram for synthesis of biotin-SS-MTX. Abbreviations: EDC, 1-ethyl-3-(3(dimethylamino)propyl)carbodiimide; NHS, N-hydroxysuccinimide; MTX-NHS, NHS ester of MTX; DAP, diaminopentane, r.t., room temperature; ppt. w/, precipitation with. The detailed synthetic procedure is described in (12).
could be hydrolyzed by carboxypeptidase G 2 (CP-G2), it was considered to be predominantly the L,~/-isomer. This conclusion is based upon two earlier observations (a) that MTX is racemized to the L- and D-enantiomers upon treatment with coupling reagents; and (b) that, in glutamate-substituted MTX derivatives, only L-isomers with a free et-COOH are subject to hydrolysis by CP-G 2 (14). Purity and authenticity of M T X - D A P and biotin-SS-MTX were established by TLC, HPLC, elemental analysis, absorbance spectra and mass spectra. Biotin-SS-folate was also synthesized by the procedure outlined in Figure 2, except that MTX was replaced by folate (12). Both biotin-SS-MTX and biotin-SS-folate contain the disulfide linkage which was essential for purification of the folate transport proteins (see below). A related compound, biotin-MTX, was also prepared by the route shown in Figure 2, except that NHSS--SS-biotin was replaced by NHS-biotin. This probe, with a non-dissociable linker, will be useful for electron microscopy of the folate transport protein in cells labeled with biotin-MTX, followed by treatment with a gold-streptavidin conjugate.
FIG. 3. Light microscopy of parental L1210 cells labeled with biotin-MTX and then exposed to streptavidin-agarose beads. The NHSS ester of biotin-MTX was prepared by conversion of biotin-MTX to the acid form (as described previously for F-MTX (9)) followed by treatment (3.5 hr) with EDC and NHS in anhydrous DMSO; the concentration was determined spectrophotometrically. L1210 cells at mid-log phase (8 x 106) were washed twice with 160 mM Hepes buffer (pH 7.4) containing 2 mM MgCI 2 and suspended in 0.2 ml of the same buffer containing 5 ~tM of the freshly-prepared NHSS ester of biotin-MTX. After 5 min, the cells were washed extensively with the same buffer. Streptavidin-agarose beads (30/~1 bed vol; 100-200/~m diameter) that had been washed twice with the above buffer were added, and after standing at 4°C for 5 min, an aliquot (3/~1) was applied to a microscope slide and examined by light microscopy. Panels A and B, labeled cells exposed to beads; C, labeled cells treated with excess streptavidin prior to exposure t o beads; D, beads treated with excess biotin prior to exposure to labeled cells. Magnification, A, C, D (xl00); B (x250).
FOLATE TRANSPORT SYSTEMS IN LI210 CELLS ABILITY
OF BIOTIN
PROBES TO LABEL PROTEINS
FOLATE
7
TRANSPORT
Probes that are suitable for labeling proteins must have two characteristics: (a) structural modification of the parent compound must not impair its binding to the target; and (b) covalent attachment must occur upon probe-target interaction. Previous studies in this laboratory (9) had shown that derivatization of MTX (on the V-position) by fluorescein-DAP did not appreciably decrease the affinity of L1210 or L. casei folate transporters for the drug (as measured by K i values for non-activated ~/-F-MTX as an inhibitor of [3H]MTX transport), and covalent attachment was evident from the ability of probe-transporter complexes to survive SDS-PAGE. Accordingly, it was assumed that the biotin derivatives of MTX and folate would behave in a similar manner. When tested as an inhibitor of [3H]MTX transport in parental L1210 cells, biotin-SS-MTX had an inhibition constant (Ki) of 0.5/zM, which is comparable to the K t value of 1.6/./,M for [3H]MTX. Thus, addition of the biotin group (via a 15-atom spacer) does not hinder the drug from interacting with the transporter. The availability of a free carboxyl on the glutamate moiety should allow the probe to be activated (by conversion to the NHSS-ester) and attached covalently to the transporter. To extend these findings, a simple visual experiment was performed. Parental L1210 cells, which contain only the /xM folate transporter, were treated with the NHS-ester of biotin-MTX, washed extensively, and exposed to streptavidin-conjugated agarose beads. As seen in Figure 3 (panel A), cells labeled with biotin-MTX were readily adsorbed onto the beads. At higher magnification (panel B), the tight packing of cells on the bead surface is evident. Specificity of the process was demonstrated by two controls: (1) labeled cells were treated with excess streptavidin before exposure to the beads (panel C); and (2) beads were treated with excess biotin prior to exposure to the labeled cells (panel D). Another control (not shown) involved treatment of cells with the NHS-ester of biotin-MTX in the presence of excess MTX; under these conditions, only a few cells were attached to the beads. Thus, it appears that the NHS-ester of biotin-MTX positions itself in the active site of the transporter and a nucleophilic group (e.g., lysine or histidine) is available for displacement of the NHS from the ot-carboxyl of the probe concomitant with formation of a covalent bond. AFFINITY
PURIFICATION
OF L1210 FOLATE PROTEINS
TRANSPORT
The strategy underlying this procedure is outlined in Figure 4. Parental L1210 cells were treated with the, NHSS-ester of biotin-SS-MTX to label the folate transport protein (ftp), and sucrose gradient centrifugation was used to
8
~
J. FAN, et al.
NHSS-ester o)~
Detergent Extracti on
Biotln-SS-MTX
TX-SS-@
L1210Cells
•
TX-SS-~)
/Streptsvldin Agarose Beads Dithiothreitol
FIG. 4. Flow diagram for purification of L1210 folate transport proteins. Abbreviations: ftp, folate transport protein; B, biotin; and SA, streptavidin. Other abbreviations are given in the text.
isolate a plasma membrane fraction. The membranes were treated with 1.2% 3-[(3-cholamidopropyl)dimethylamino]-l-propanesulfonate (CHAPS), and the detergent extract was then exposed to streptavidin agarose beads. The beads were washed extensively to remove extraneous proteins and then treated with dithiothreitol (DTI'). SDS-PAGE gels of the detergent extract before and after exposure to the streptavidin beads (not shown) appeared to be identical, indicating that little, if any, non-specific adsorption of membrane proteins onto the beads had occurred during this procedure. Examination of the DTT eluate of the beads, however, revealed that a single protein with an apparent molecular weight of ca. 43 kDa (Fig. 5, lane a) had been adsorbed from the detergent extract and subsequently released by the dithiol. This entity is considered to be the L1210 p.M folate transporter, based upon the labeling and isolation procedure and the agreement between its molecular weight with values obtained earlier in this laboratory (9, 13) and by other investigators who utilized the NHS-ester of [3H]aminopterin (10) or a photo-activated, 125I-labeled analog of MTX (11) to label the protein in situ, followed by radioautography of the membrane extracts. It should be noted that all steps in the above procedure, beginning with the isolation of plasma membranes, were conducted in the presence of multiple protease inhibitors (pepstatin A, aprotinin, leupeptin and phenylmethylsulfonyl fluoride). When the isolation procedure was repeated with the protease inhibitors omitted during one step, the 43 kDa band was accompanied by a less intense 36 kDa band (Fig. 5, lane B). The latter protein, which has been observed previously by Henderson (15), appears to be a proteolytie fragment of the/zM transporter. Cleavage of the 43 kDa protein (by an unidentified
i~iiiiiiii~!iiiii'!iiiii~!~!~i!ii~!!!!i~!iiii~j!i!!!!!i!iiiiii~iiii~ i~,¸~i~ ~
66--~
45---~ 36--~
24--~
20"~
A
B
FIG. 5. SDS-PAGE of purified/,I.M folate transport protein from parental L1210 cells. (A) ~M transporter (5/zg) purified in the presence of protease inhibitors; (B) same except that protease inhibitors were omitted. Stain, Poinceau S, after transfer of the protein to polyvinylidene difluoride (PVDF) membrane.
66"-~ 45.-,.-
-,-43 --~--39
36--~ 29.-~
20--,-14--~
FIG. 6. SDS-PAGE of purified nM and //.M folate transport proteins from JF subline of L1210 cells. Approximately 10 txg protein were applied to the gel. Stain and transfer to PVDF membrane as in Figure 5.
FOLATE TRANSPORT SYSTEMS IN LI210 CELLS protease located presumably in L1210 plasma membranes) appears to be quite specific; no other fragments were seen on the electrophoretogram. The physiological significance of this proteolytic process, if any, remains to be investigated. A similar procedure was utilized for isolation of the L1210 nM folate transporter, except that biotin-SS-folate was substituted for biotinSS-MTX, and parental L1210 cells were replaced by a subline (JF) up-regulated for the nM transporter. The latter cells were obtained by propagating parental cells on very low (i.e., nanomolar) concentrations of folate; after six months on this regimen, the level of the nM transporter in the JF clone had increased from essentially zero to ca. 6 pmoles/106 cells. In this instance, examination by S D S - P A G E of the DTI" eluate of the streptavidin beads revealed a major diffuse band at ca. 39 kDa and a minor band at 43 kDa (Fig. 6). The former was judged to be the nM transporter, and the intensity of its staining with Coomassie Blue relative to that of the 43 kDa entity was consistent with the 60-fold difference in amounts of the nM and/~M folate transporters in the up-regulated subline (cf. Table 1). This experiment clearly demonstrated that the/~M transporter had not disappeared during the slow acquisition of the nM counterpart, a finding that would have been
TABLE 2. AMINO ACID COMPOSITION* OF FOLATE TRANSPORT PROTEINS Transport protein Amino acid Ala Arg Asx Cys Glx Gly His
L1210~M
L1210 nM
8.1 5.6 10.1 --~ 12.2 9.4 3.0
5.9 6.0 11.3 -13.8 6.4 4.9
Ile Leu Lys Met Phe Pro Ser Thr
Trp Tyr Val *Percent of total amino acids. tFrom Henderson et al. (2). .~Not determined.
L. caseit
9.9 3.6 6.3 0.1 4.1 6.9 1.2
4.8
3.1
8.1
7.1 5.6 0.3 3.9 4.7 8.2 6.8 -5.3 4.9
5.5 6.7 0.2 4.5 5.2 10.2 6.1 -7.1 3.0
11.6 3.7 8.0 5.8 5.4 6.4 6.7 4.7 2.2 5.6
10
J. FAN, et al.
difficult to reveal by conventional kinetic measurements. Also of interest is the observation that omission of the protease inhibitors during the isolation procedure did not lead to degradation of the nM transporter (not shown).
CHARACTERIZATION
OF L1210 FOLATE
TRANSPORT
PROTEINS
Amino acid compositions of the L1210 ~M and nM folate transport proteins, along with that of the L. casei transporter, are given in Table 2. The/xM and nM transport proteins are not substantially different, although the former has a slightly higher proportion of hydrophobic amino acids. T h e L. casei transporter, however, is considerably more hydrophobic. Attempts to obtain N-terminal sequences of the transport proteins (and the 36 kDa fragment from the ~M transporter), as a prelude to synthesizing oligonucleotide probes for cloning experiments, were unsuccessful. Since these results appear to be due to the presence of blocked N-termini, it will be necessary to acquire sequence information from peptides obtained via enzymatic digestion or CNBr-treatment of the proteins. The presence of carbohydrate in the L1210 folate transporters was examined by an assay in which each protein was treated with peptide:N-
[ Ethanolamine I
Extracellular Space Plasma Membrane
FIG. 7. Schematic representation of the linkage of a protein to the cell membrane via a glycosylphosphatidylinositol (GPI) "tail". Abbreviations, P, phosphate; I, inositoi; PI-PLC, phosphatidylinositol-specific phospholipase C.
FIG. 8. Fluorescence microscopy of fluorescein-labeled parental and JF subline L1210 cells before and after treatment with PI-PLC. Cells were labeled as described in the legend to Figure 3, except that fluorescein-folate was used in the place of biotin-MTX. The probe was used directly (without prior conversion to the NHSS ester). A, labeled JF subline cells; B, same, after treatment with PI-PLC (1.5 units; 2 hr; 37°C); labeled parental L1210 cells; and D, same as panel C except treated with PI-PLC.
FOLATE TRANSPORTSYSTEMSIN LI210 CELLS
11
glycosidase F (N-glycanase), an enzyme that removes asparagine-linked sugars from proteins. The/xM transporter was unaffected by this procedure, as evidenced by retention of the 43 kDa band on SDS-PAGE gels and the lack of any new bands (data not shown). In contrast, the band corresponding to the 39 kDa transporter disappeared, and a new band (masked, unfortunately, by the glycanase itself) at 32 kDa appeared (12). A similar susceptibility to glycanase has been observed previously with the nM folate transporter from human KB cells (3) and human placenta (5). Anchorage of the /zM and nM folate transporters to the plasma membrane of L1210 cells was examined by fluorescence microscopy with the aid of fluorescein-MTX (or -folate) and phosphatidylinositol-specific phospholipase C (PI-PLC). This enzyme hydrolyzes glycosylphosphatidylinositol (GPI) "tails" at the glycerol-phosphate linkage (as shown by the dashed line in Fig. 7), and therefore releases the protein from the membrane. Accordingly, L1210 cells up-regulated for the nM transporter were labeled non-covalently with a fluorescein derivative of folic acid (synthesized by a procedure similar to that used for F-MTX (9)) and then treated with PI-PLC (Fig. 8, panels A and B). Complete loss of the cell fluorescence indicated that the nM transporter had been released from the membrane. In control experiments, no diminution in fluorescence was observed when enzyme was omitted, and labeled cells treated with enzyme did not become fluorescent upon re-exposure to the probe (data not shown). In contrast, when parental cells (containing the ~M transporter) were labeled with F-MTX and treated with PI-PLC, the fluorescence intensity remained constant (Fig. 8, panels C and D). It should also be noted that Figure 8 (compare panels A and C) provides visual evidence that the nM transporter in the JF subline is present at a much higher level than its/xM counterpart in the parental line. These results indicate that the L1210/.tM folate transporter is a 43 kDa integral membrane protein and does not contain asparagine-linked carbohydrate. In contrast, the /zM transporter from K562 cells (16), and possibly from various other human malignant cell lines as well, is heavily glycosylated and has a much higher molecular weight (ca. 80 kDa). Since the various murine and human/~M folate transporters have very similar kinetic properties (16), the rationale for this discrepancy in carbohydrate content and molecular weight is not apparent. The L1210 nM folate transporter is a 39 kDa glycosylated protein anchored exofacially on the plasma membrane by a GPI tail. In these properties it closely resembles its counterparts from human sources (e.g., KB cells (3), Caco cells (4) and placenta (5)) and from MA-104 monkey kidney cells (4). The nM transporters from KB cells, MA-104 cells and placenta have very similar amino acid sequences, and it is likely that the L1210 nM protein will not be too different.
12
J. FAN, et al. FUTURE DIRECTIONS
Based upon the results described above, it seems likely that biotin-SSMTX and biotin-SS-folate probes could be used to isolate folate transport proteins from other sources. Proteins responsible for folate efflux might also be identified by using the probes with membrane fragments, in which the inner membrane surface would be accessible. The general procedure might also be adapted to the identification and isolation of folate-dependent enzymes from cytoplasmic extracts. In both instances, however, functionality of the protein would not be demonstrable because of the covalent bond between the probe and the binding site. Future work should be directed toward development of a structure that would allow for reversible attachment of the probe. Another use of the biotin derivatives of MTX and folate is to visualize the folate transporters in individual cells. This has been accomplished by treating L1210 cells with these probes followed by a fluorescein conjugate of streptavidin. Results similar to those obtained with the fluorescein derivatives of MTX and folate (Fig. 8) are obtained (data not shown). Cross-labeling of the transporters (when both are present in the same cell) can be avoided by using: (a) activated biotin-MTX in the presence of a large excess of folate when only the/~M transporter is to be visualized; and (b) non-activated biotin-folate for exclusive labeling of the nM transporter. The use of fluorescein-streptavidin and rhodamine-streptavidin should allow color-specific labeling of the individual transporters to be obtained. One can envision many additional uses for these probes as visual markers of the folate transporters. These include: (1) To follow changes, if they occur, in the levels of these transporters during cell growth or cell cycle. (2) To assess the level of folate transporters in cells from cancer patients. This may be especially important in view of the report that the /zM folate transporter disappears during the DMSO-induced differentiation (and loss of malignancy) of murine erythroleukemia cells (17). It would be of interest to determine whether the /zM transporter appears during transformation of normal cells to tumor cells. (3) To investigate the mechanism for up-regulation of the nM folate transporter when cells are adapted to grow on low levels of folate. There are at least three possible explanations of this phenomenon: (i) selection of pre-existing cells with high levels of the transporter; (ii) loss of a folate-repressor protein complex that regulates expression of the transporter gene; or (iii) amplification of the transporter gene during growth perturbations caused by folate limitation. (4) To delineate the mechanism by which the nM transporter, which appears to be localized entirely outside of the membrane, translocates bound substrates through the membrane. Kamen and his colleagues (18) have proposed an endocytotic mechanism involving
FOLATE TRANSPORTSYSTEMSIN LI210CELLS
13
caveoli instead of coated pits without lysosomal participation, followed by transfer to the H,M transporter. SUMMARY Biotin derivatives of methotrexate (biotin-SS-MTX) and folate (biotin-SS-folate), in which the functional components are joined by a dissociable disulfide-containing spacer, have been synthesized, purified by DEAE-Trisacryl chromatography, and characterized by HPLC, elemental analysis and mass spectrometry. These compounds provide a convenient means for the single-step purification of the folate transporters from L1210 cells. Parental L1210 murine leukemia cells, which contain only the /.,M transporter (the reduced folate/MTX transport protein) were treated with the N-hydroxysulfosuccinimide ester of biotin-SS-MTX, and a detergent extract of the plasma membranes was exposed to streptavidin-agarose beads to adsorb the labeled protein. Dithiothreitol cleavage of the disulfide linkage released the transporter, which migrated as a well-defined component (43 kDa) on SDS-PAGE gels; no other proteins were present. An L1210 subline (JF), obtained by adapting cells to grow on nanomolar concentrations of folate, contains both the /zM transporter and the nM transporter (high-affinity folate binding protein). When these cells were treated with the N-hydroxysulfosuccimide ester of biotin-SS-folate and processed as described above, analysis on SDS-PAGE gels revealed the presence of two proteins, the /~M transporter (43 kDa) and the nM transporter (39 kDa). Both transporters were characterized with respect to amino acid content; blocked N-termini precluded Edman sequencing. Treatment of the nM transporter with peptide:N-glycosidase F produced a smaller component (32 kDa); the p.M transporter, conversely, was unchanged by this procedure. When the /zM transporter in parental L1210 cells was labeled with fluorescein-MTX and then treated with phosphoinositol-specific phospholipase C (PI-PLC), no change in fluorescence was detected. Alternatively, when the nM transporter in the JF subline-was labeled with fluorescein-folate and then treated with PI-PLC, complete loss of fluorescence was observed. These results indicate that the L1210/.tM transporter is a non-glycosylated, integral membrane protein, while its nM counterpart is heavily glycosylated and anchored exofacially to the membrane by a glycosylphosphatidylinositol component. ACKNOWLEDGEMENTS This investigation was supported by an Outstanding Investigator Grant (CA 39836) from the National Cancer Institute and a Grant (CH-31) from
14
J. FAN, et al.
the American Cancer Society. The authors are indebted to Ms Susan Burke for expert assistance in the preparation of the manuscript. REFERENCES 1. D. KESSEL, T. C. HALL, D. ROBERTS and I. WODINSKY, Uptake as a determinant of methotrexate response in mouse leukemias, Science 150, 752-754 (1965). 2. G . B . HENDERSON, E. M. ZEVELY and F. M. HUENNEKENS, Purification and properties of a membrane-associated folate-binding protein from Lactobacillus casei, J. Biol Chem. 252, 3760-3765 (1977). 3. P . C . ELWOOD, Molecular cloning and characterization of the human folate-binding protein cDNA from placenta and malignant tissue culture (KB) cells, J. Biol. Chem. 264, 14893-14901 (1989). 4. S. W. LACEY, J. M. SANDERS, K. G. ROTHBERT, R. G. W. ANDERSON and B. A. KAMEN, Complementary DNA for the folate binding protein correctly predicts anchoring to the membrane by glycosyl-phosphatidylinositol, J. Clin. Invest. 84, 715-720 (1989). 5. M. RATNAM, H. M A R Q U A R D T , J. L. DUHRING and J. H. FREISHEIM. Homologous membrane folate binding proteins in human placenta: cloning and sequence of a cDNA, Biochemistry 28, 8249-8254 (1989). 6. G. B. HENDERSON, Transport of folate compounds into cells, pp. 207-250 in Nutritional, Pharmacologic and Physiologic Aspects of Folates and Pterins. (R. L. BLAKLEY and M. V. WHITEHEAD, eds.) John Wiley and Sons, New York (1986). 7. G. B. HENDERSON, J. M. TSUJI and H. P. KUMAR, Mediated uptake of folate by a high-affinity binding protein in sublines of L1210 cells adapted to nanomolar concentrations of folate, J. Membrane Biol. 101, 247-258 (1988). 8. G. JANSEN, G. R. WESTERHOF, 1. KATHMANN, B. C. RADEMAKER, G. RIJKSEN and J. H. SCHORNAGEL, Identification of a membrane-associated folate-binding protein in human leukemic CCRF-CEM cells with transport-related methotrexate resistance, Cancer Res. 49, 2455-2459 (1989). 9. J. FAN, L. E. POPE, K. S. VITOLS and F. M. HUENNEKENS, Affinity labeling of folate transport proteins with the N-hydroxysuccinimide ester of the ~-isomer of fiuorescein methotrexate, Biochemistry 30, 4573-4580 (1991). 10. C.-H. YANG, F. M. SIROTNAK and L. S. MINES, Further studies on a novel class of genetic variants of the LI210 cell with increased folate analogue transport inward, J. Biol. Chem. 263, 9703-9709 (1988). 11. E.M. PRICE and J. H. FREISHEIM, Photoaffinity analogues of methotrexate as folate antagonist binding probes. 2. Transport studies, photoaffinity labeling, and identification of the membrane carrier protein for methotrexate for murine L1210 cells, Biochemistry 26, 4757-4763 (1987). 12. J. FAN, K. S. VITOLS and F. M. HUENNEKENS, Biotin derivatives of methotrexate and folate: synthesis and utilization for affinity purification of two membrane-associated folate transporters from L1210 cells, J. Biol. Chem. 266, 14862-14865 (1991). 13. J. FAN, L. E. POPE, K. S. VITOLS and F. M. HUENNEKENS, Visualization of folate transport proteins by covalent labeling with fluorescein methotrexate, Advan. Enzyme ReguL 30, 3-12 (1990). 14. U. KUEFNER, A. ESSWEIN, U. LOHRMANN, Y. MONTEJANO, K. S. VITOLS and F. M. HUENNEKENS, Occurrence and significance of diastereomers of methotrexate ct-peptides, Biochemistry 29, 10540-10545 (1990). 15. G. B. HENDERSON and E. M. ZEVELY, Affinity labeling of the 5methyltetrahydrofolate/methotrexate transport protein of L1210 cells by treatment with an N-hydroxysuccinimide ester of [3H]methotrexate, J. Biol. Chem. 2,59, 4558-4562 (1984). 16. L . H . MATHERLY, C. A. CZAJKOWSKI and S. M. ANGELES, Identification of a highly glycosylated methotrexate membrane carrier in K562 human erythroleukemia
FOLATE TRANSPORT SYSTEMS IN LI210 CELLS
15
cells up-regulated for tetrahydrofolate cofactor and methotrexate transport, Cancer Res. 51, 3420-3426 (1991). 17. R. E. CORIN, H. C. HASPEL and M. SONENBERG, Transport of the folate compound methotrexate decreases during differentiation of murine erythroleukemia cells, J. Biol. Chem. 259, 206-211 (1984). 18. B. A. KAMEN, A. K. SMITH and R. G. W. ANDERSON, The folate receptor works in tandem with a probenecid-sensitive carrier in MA104 cells in vitro, J. Clin. Invest. 87, 1442-1449 (1991).
0065-2571/92/$15.00 © 1992 Pergamon Pressplc.
MULTIPLE FOLATETRANSPORT SYSTEMS IN L1210CELLS J. FAN, K. S. VITOLS and F. M. HUENNEKENS Division of Biochemistry, Department of Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, CA 92037
INTRODUCTION: TRANSPORT AS A CRITICAL PARAMETER THE CHEMOTHERAPEUTIC EFFICACY OF METHOTREXATE
IN
Methotrexate (MTX)~, the 4-amino-10-methyl analog of folic acid, was one of the first agents shown to be effective in cancer chemotherapy, and it continues to be used extensively for the treatment of choriocarcinoma, acute lymphocytic leukemia, and, in combination with other agents, for maintenance of drug-induced remissions. Extensive information is available for MTX with regard to its transport into cells, metabolic conversion to polyglutamate forms, and inhibition of dihydrofolate reductase (DHFR; EC 1.5.1.3). Resistance to MTX, the result of defective transport or polyglutamylation, of increased levels of the target enzyme DHFR, or of mutations in DHFR, has been documented in cellular model systems and in patients undergoing drug therapy. In addition to its intrinsic value, this large body of information about MTX has provided valuable guidance for parallel studies with other drugs. Recognition of transport as a critical parameter in the chemotherapeutic efficacy of MTX began with the classic report of Kessel, Hall and Roberts in 1965 (1). These investigators, using a series of murine tumor cell lines, found that a linear relationship existed between the rate of uptake of M T X in vitro and the ability of the drug to kill these cells in vivo. Subsequent studies in a number of laboratories, conducted primarily with L1210 murine leukemia cells, demonstrated that MTX was taken up by an active transport system whose normal substrates were folate compounds, and the substrate specificity and kinetic characteristics of folate transport systems in various cells and tissues were delineated. This work, which utilized intact cells, was able to provide only limited information about components, mechanism, or regulation of the transport systems. In 1977, however, a membrane-associated folate binding protein, shown to be essential for folate transport, was purified to homogeneity from Lactobacillus casei (2) and, more recently, folate transporters from KB cells (3), MA-104 cells (4) and human placenta (5) have been isolated and characterized.
4
J. FAN, et al. G E N E R A L P R O P E R T I E S OF F O L A T E T R A N S P O R T SYSTEMS OF L1210 CELLS
An unexpected complication in folate transport has emerged from the evidence that certain eukaryotic cells can express two different systems for this purpose. L1210 cells, for example, when propagated in vitro on the usual laboratory media containing m i c r o m o l a r concentrations of folate, express a system whose primary substrate is 5-methyltetrahydrofolate (K t ~ 1 /~M). 5-Formyltetrahydrofolate (folinate) and M T X are also good substrates for this system, as indicated by their K t values of ca. 5/~M, but folate (K t > 100/~M) is a poor substrate (reviewed in (6)). Alternatively, growth of L1210 cells on nanomolar concentrations of folate or folinate leads to the gradual expression of a separate system (7) in which folate is the primary substrate (K t < 1 nM). 5-Methyltetrahydrofolate, 5-formyltetrahydrofolate and M T X are transported less efficiently, although their K t values are still in the nanomolar range. Based upon their kinetic characteristics, these will be referred to hereafter as the L1210 "nM" and "/~M" folate transport systems. Other investigators have preferred the term "reduced folate/MTX" for the latter system. Still others have designated the two systems as "high affinity/low capacity" (nM) and "low affinity/high capacity" (/~M). Properties of these folate transport systems are summarized in Table 1. Although the /~M transporter is expressed in a much smaller amount (ca. 60-fold) than the nM transport protein, the higher rate of turnover (ca. 7000-fold) results in a 100-fold higher Vmax. Thus, the overall efficiency of the ~M
TABLE 1. FOLATE TRANSPORT SYSTEMS IN L1210CELLS Property
~M System
nM System
Kt Vmax,pmol/min/106cells Turnover number*
mFH4 > fFH4 -- MTX > F 1 ~tM 5 ttM 100/zM 0.5 5
F > mFH4 = FH4 > MTX <1 nM 10 nM 100 nM 5 x 10-3 7 x 10-4
Amount, pmol/106 cells number/cell Molecular weight, kDA
0. l 6 x 104 43
6 3.6 x 106 39 Yes Yes
Substrate specificity
CHOt GP.t Energy source Mechanism Regulation
No No Anion gradient Anion exchange cAMP
? Endocytosis ? Tandem-/tM transporter ? ?
*Molecules/min/transporter. tAsparagine-linked carbohydrate detected by treatment with peptide:N-glycosidaseF. ~Glycosylphosphatidylinositol.
FOLATE TRANSPORTSYSTEMSIN LI210CELLS transport system is much greater than that of its nM counterpart. A similar coexistence of the nM and/~M folate transport systems has been observed in CCRF-CEM human leukemia cells (8), and the two transport systems have been characterized individually in other eukaryotic cells. Prior to the present study, relatively little was known about the molecular properties of the L1210/£M and nM folate transporters, although labeling experiments had indicated that the molecular weight of the former was 43--45 kDa (9-11). This paucity of information was largely due to the low level of the/~M transporter and to the difficulty in obtaining sublines upregulated for expression of either transporter. The present investigation was undertaken, therefore, to develop a rapid and efficient procedure for isolating these proteins. Biotin derivatives of MTX and folate containing a dissociable S-S linker have been synthesized and used to label the transporters in intact cells (12). Streptavidin-agarose beads are employed to collect the labeled proteins from detergent extracts of the isolated cell membranes, and the proteins are released by reduction of the S--S bond in the probes. SYNTHESIS O F B I O T I N D E R I V A T I V E S FOLATE
OFMETHOTREXATE
AND
The prototype compound in this series was biotin-SS-MTX, whose structure is shown in Figure 1. The structure is similar to fluorescein-MTX (9, 13), except that the spacer group is lengthened by the addition of a dissociable disulfide bond and fluorescein is replaced by biotin. The synthetic procedure for biotin-SS-MTX is outlined in Figure 2. MTX was converted to the N-hydroxysuccinimide (NHS)-ester, and the latter was reacted with diaminopentane (DAP). The product MTX-DAP, after purification by ion-exchange chromatography, was linked to the commercially available N-hydroxysulfosuccinimide (NHSS)-ester of a biotin compound containing a disulfide group. This sequence was modified from that used for the preparation of F-MTX (viz., FITC + DAP ---> FDAP ---> F-MTX) because of the limited availability of NHSS-SS-biotin and the non-chromophoric nature of biotin. Biotin-SS-MTX was obtained in an overall yield of ca. 50%, and since only 50-60% of the compound
","~" ,y% NHI
~
coo. !
,~, ~-"-(¢"')'-"- .c. g-Cc.,),-s-s -(c.,),-..- c-Co.,).
FIG. I. Structure of biotin-SS-MTX.
J. FAN, et al.
MTX 0 x) (a) EDC, lh, r.t. (1.2x) (b) NHS, 4h, r.t. (1.2x)
DAP
(lOx)
+
MTX-NHS Ox)
l
(a) 3h, r.t. (b) ppt. w/acetone (c) DEAE-TrlIacryl (H,O)
MTX-DAP
+
NHSS-SS-Biotin
(Ix)
(lx)
I (a) 6h, r.t, I (b) ppt. w/acetone ] (c) DEAE-TrlIecryl ~L (0.05M NH4HCO,) m
Biotin-SS-MTX FIG. 2. Flow diagram for synthesis of biotin-SS-MTX. Abbreviations: EDC, 1-ethyl-3-(3(dimethylamino)propyl)carbodiimide; NHS, N-hydroxysuccinimide; MTX-NHS, NHS ester of MTX; DAP, diaminopentane, r.t., room temperature; ppt. w/, precipitation with. The detailed synthetic procedure is described in (12).
could be hydrolyzed by carboxypeptidase G 2 (CP-G2), it was considered to be predominantly the L,~/-isomer. This conclusion is based upon two earlier observations (a) that MTX is racemized to the L- and D-enantiomers upon treatment with coupling reagents; and (b) that, in glutamate-substituted MTX derivatives, only L-isomers with a free et-COOH are subject to hydrolysis by CP-G 2 (14). Purity and authenticity of M T X - D A P and biotin-SS-MTX were established by TLC, HPLC, elemental analysis, absorbance spectra and mass spectra. Biotin-SS-folate was also synthesized by the procedure outlined in Figure 2, except that MTX was replaced by folate (12). Both biotin-SS-MTX and biotin-SS-folate contain the disulfide linkage which was essential for purification of the folate transport proteins (see below). A related compound, biotin-MTX, was also prepared by the route shown in Figure 2, except that NHSS--SS-biotin was replaced by NHS-biotin. This probe, with a non-dissociable linker, will be useful for electron microscopy of the folate transport protein in cells labeled with biotin-MTX, followed by treatment with a gold-streptavidin conjugate.
FIG. 3. Light microscopy of parental L1210 cells labeled with biotin-MTX and then exposed to streptavidin-agarose beads. The NHSS ester of biotin-MTX was prepared by conversion of biotin-MTX to the acid form (as described previously for F-MTX (9)) followed by treatment (3.5 hr) with EDC and NHS in anhydrous DMSO; the concentration was determined spectrophotometrically. L1210 cells at mid-log phase (8 x 106) were washed twice with 160 mM Hepes buffer (pH 7.4) containing 2 mM MgCI 2 and suspended in 0.2 ml of the same buffer containing 5 ~tM of the freshly-prepared NHSS ester of biotin-MTX. After 5 min, the cells were washed extensively with the same buffer. Streptavidin-agarose beads (30/~1 bed vol; 100-200/~m diameter) that had been washed twice with the above buffer were added, and after standing at 4°C for 5 min, an aliquot (3/~1) was applied to a microscope slide and examined by light microscopy. Panels A and B, labeled cells exposed to beads; C, labeled cells treated with excess streptavidin prior to exposure t o beads; D, beads treated with excess biotin prior to exposure to labeled cells. Magnification, A, C, D (xl00); B (x250).
FOLATE TRANSPORT SYSTEMS IN LI210 CELLS ABILITY
OF BIOTIN
PROBES TO LABEL PROTEINS
FOLATE
7
TRANSPORT
Probes that are suitable for labeling proteins must have two characteristics: (a) structural modification of the parent compound must not impair its binding to the target; and (b) covalent attachment must occur upon probe-target interaction. Previous studies in this laboratory (9) had shown that derivatization of MTX (on the V-position) by fluorescein-DAP did not appreciably decrease the affinity of L1210 or L. casei folate transporters for the drug (as measured by K i values for non-activated ~/-F-MTX as an inhibitor of [3H]MTX transport), and covalent attachment was evident from the ability of probe-transporter complexes to survive SDS-PAGE. Accordingly, it was assumed that the biotin derivatives of MTX and folate would behave in a similar manner. When tested as an inhibitor of [3H]MTX transport in parental L1210 cells, biotin-SS-MTX had an inhibition constant (Ki) of 0.5/zM, which is comparable to the K t value of 1.6/./,M for [3H]MTX. Thus, addition of the biotin group (via a 15-atom spacer) does not hinder the drug from interacting with the transporter. The availability of a free carboxyl on the glutamate moiety should allow the probe to be activated (by conversion to the NHSS-ester) and attached covalently to the transporter. To extend these findings, a simple visual experiment was performed. Parental L1210 cells, which contain only the /xM folate transporter, were treated with the NHS-ester of biotin-MTX, washed extensively, and exposed to streptavidin-conjugated agarose beads. As seen in Figure 3 (panel A), cells labeled with biotin-MTX were readily adsorbed onto the beads. At higher magnification (panel B), the tight packing of cells on the bead surface is evident. Specificity of the process was demonstrated by two controls: (1) labeled cells were treated with excess streptavidin before exposure to the beads (panel C); and (2) beads were treated with excess biotin prior to exposure to the labeled cells (panel D). Another control (not shown) involved treatment of cells with the NHS-ester of biotin-MTX in the presence of excess MTX; under these conditions, only a few cells were attached to the beads. Thus, it appears that the NHS-ester of biotin-MTX positions itself in the active site of the transporter and a nucleophilic group (e.g., lysine or histidine) is available for displacement of the NHS from the ot-carboxyl of the probe concomitant with formation of a covalent bond. AFFINITY
PURIFICATION
OF L1210 FOLATE PROTEINS
TRANSPORT
The strategy underlying this procedure is outlined in Figure 4. Parental L1210 cells were treated with the, NHSS-ester of biotin-SS-MTX to label the folate transport protein (ftp), and sucrose gradient centrifugation was used to
8
~
J. FAN, et al.
NHSS-ester o)~
Detergent Extracti on
Biotln-SS-MTX
TX-SS-@
L1210Cells
•
TX-SS-~)
/Streptsvldin Agarose Beads Dithiothreitol
FIG. 4. Flow diagram for purification of L1210 folate transport proteins. Abbreviations: ftp, folate transport protein; B, biotin; and SA, streptavidin. Other abbreviations are given in the text.
isolate a plasma membrane fraction. The membranes were treated with 1.2% 3-[(3-cholamidopropyl)dimethylamino]-l-propanesulfonate (CHAPS), and the detergent extract was then exposed to streptavidin agarose beads. The beads were washed extensively to remove extraneous proteins and then treated with dithiothreitol (DTI'). SDS-PAGE gels of the detergent extract before and after exposure to the streptavidin beads (not shown) appeared to be identical, indicating that little, if any, non-specific adsorption of membrane proteins onto the beads had occurred during this procedure. Examination of the DTT eluate of the beads, however, revealed that a single protein with an apparent molecular weight of ca. 43 kDa (Fig. 5, lane a) had been adsorbed from the detergent extract and subsequently released by the dithiol. This entity is considered to be the L1210 p.M folate transporter, based upon the labeling and isolation procedure and the agreement between its molecular weight with values obtained earlier in this laboratory (9, 13) and by other investigators who utilized the NHS-ester of [3H]aminopterin (10) or a photo-activated, 125I-labeled analog of MTX (11) to label the protein in situ, followed by radioautography of the membrane extracts. It should be noted that all steps in the above procedure, beginning with the isolation of plasma membranes, were conducted in the presence of multiple protease inhibitors (pepstatin A, aprotinin, leupeptin and phenylmethylsulfonyl fluoride). When the isolation procedure was repeated with the protease inhibitors omitted during one step, the 43 kDa band was accompanied by a less intense 36 kDa band (Fig. 5, lane B). The latter protein, which has been observed previously by Henderson (15), appears to be a proteolytie fragment of the/zM transporter. Cleavage of the 43 kDa protein (by an unidentified
i~iiiiiiii~!iiiii'!iiiii~!~!~i!ii~!!!!i~!iiii~j!i!!!!!i!iiiiii~iiii~ i~,¸~i~ ~
66--~
45---~ 36--~
24--~
20"~
A
B
FIG. 5. SDS-PAGE of purified/,I.M folate transport protein from parental L1210 cells. (A) ~M transporter (5/zg) purified in the presence of protease inhibitors; (B) same except that protease inhibitors were omitted. Stain, Poinceau S, after transfer of the protein to polyvinylidene difluoride (PVDF) membrane.
66"-~ 45.-,.-
-,-43 --~--39
36--~ 29.-~
20--,-14--~
FIG. 6. SDS-PAGE of purified nM and //.M folate transport proteins from JF subline of L1210 cells. Approximately 10 txg protein were applied to the gel. Stain and transfer to PVDF membrane as in Figure 5.
FOLATE TRANSPORT SYSTEMS IN LI210 CELLS protease located presumably in L1210 plasma membranes) appears to be quite specific; no other fragments were seen on the electrophoretogram. The physiological significance of this proteolytic process, if any, remains to be investigated. A similar procedure was utilized for isolation of the L1210 nM folate transporter, except that biotin-SS-folate was substituted for biotinSS-MTX, and parental L1210 cells were replaced by a subline (JF) up-regulated for the nM transporter. The latter cells were obtained by propagating parental cells on very low (i.e., nanomolar) concentrations of folate; after six months on this regimen, the level of the nM transporter in the JF clone had increased from essentially zero to ca. 6 pmoles/106 cells. In this instance, examination by S D S - P A G E of the DTI" eluate of the streptavidin beads revealed a major diffuse band at ca. 39 kDa and a minor band at 43 kDa (Fig. 6). The former was judged to be the nM transporter, and the intensity of its staining with Coomassie Blue relative to that of the 43 kDa entity was consistent with the 60-fold difference in amounts of the nM and/~M folate transporters in the up-regulated subline (cf. Table 1). This experiment clearly demonstrated that the/~M transporter had not disappeared during the slow acquisition of the nM counterpart, a finding that would have been
TABLE 2. AMINO ACID COMPOSITION* OF FOLATE TRANSPORT PROTEINS Transport protein Amino acid Ala Arg Asx Cys Glx Gly His
L1210~M
L1210 nM
8.1 5.6 10.1 --~ 12.2 9.4 3.0
5.9 6.0 11.3 -13.8 6.4 4.9
Ile Leu Lys Met Phe Pro Ser Thr
Trp Tyr Val *Percent of total amino acids. tFrom Henderson et al. (2). .~Not determined.
L. caseit
9.9 3.6 6.3 0.1 4.1 6.9 1.2
4.8
3.1
8.1
7.1 5.6 0.3 3.9 4.7 8.2 6.8 -5.3 4.9
5.5 6.7 0.2 4.5 5.2 10.2 6.1 -7.1 3.0
11.6 3.7 8.0 5.8 5.4 6.4 6.7 4.7 2.2 5.6
10
J. FAN, et al.
difficult to reveal by conventional kinetic measurements. Also of interest is the observation that omission of the protease inhibitors during the isolation procedure did not lead to degradation of the nM transporter (not shown).
CHARACTERIZATION
OF L1210 FOLATE
TRANSPORT
PROTEINS
Amino acid compositions of the L1210 ~M and nM folate transport proteins, along with that of the L. casei transporter, are given in Table 2. The/xM and nM transport proteins are not substantially different, although the former has a slightly higher proportion of hydrophobic amino acids. T h e L. casei transporter, however, is considerably more hydrophobic. Attempts to obtain N-terminal sequences of the transport proteins (and the 36 kDa fragment from the ~M transporter), as a prelude to synthesizing oligonucleotide probes for cloning experiments, were unsuccessful. Since these results appear to be due to the presence of blocked N-termini, it will be necessary to acquire sequence information from peptides obtained via enzymatic digestion or CNBr-treatment of the proteins. The presence of carbohydrate in the L1210 folate transporters was examined by an assay in which each protein was treated with peptide:N-
[ Ethanolamine I
Extracellular Space Plasma Membrane
FIG. 7. Schematic representation of the linkage of a protein to the cell membrane via a glycosylphosphatidylinositol (GPI) "tail". Abbreviations, P, phosphate; I, inositoi; PI-PLC, phosphatidylinositol-specific phospholipase C.
FIG. 8. Fluorescence microscopy of fluorescein-labeled parental and JF subline L1210 cells before and after treatment with PI-PLC. Cells were labeled as described in the legend to Figure 3, except that fluorescein-folate was used in the place of biotin-MTX. The probe was used directly (without prior conversion to the NHSS ester). A, labeled JF subline cells; B, same, after treatment with PI-PLC (1.5 units; 2 hr; 37°C); labeled parental L1210 cells; and D, same as panel C except treated with PI-PLC.
FOLATE TRANSPORTSYSTEMSIN LI210 CELLS
11
glycosidase F (N-glycanase), an enzyme that removes asparagine-linked sugars from proteins. The/xM transporter was unaffected by this procedure, as evidenced by retention of the 43 kDa band on SDS-PAGE gels and the lack of any new bands (data not shown). In contrast, the band corresponding to the 39 kDa transporter disappeared, and a new band (masked, unfortunately, by the glycanase itself) at 32 kDa appeared (12). A similar susceptibility to glycanase has been observed previously with the nM folate transporter from human KB cells (3) and human placenta (5). Anchorage of the /zM and nM folate transporters to the plasma membrane of L1210 cells was examined by fluorescence microscopy with the aid of fluorescein-MTX (or -folate) and phosphatidylinositol-specific phospholipase C (PI-PLC). This enzyme hydrolyzes glycosylphosphatidylinositol (GPI) "tails" at the glycerol-phosphate linkage (as shown by the dashed line in Fig. 7), and therefore releases the protein from the membrane. Accordingly, L1210 cells up-regulated for the nM transporter were labeled non-covalently with a fluorescein derivative of folic acid (synthesized by a procedure similar to that used for F-MTX (9)) and then treated with PI-PLC (Fig. 8, panels A and B). Complete loss of the cell fluorescence indicated that the nM transporter had been released from the membrane. In control experiments, no diminution in fluorescence was observed when enzyme was omitted, and labeled cells treated with enzyme did not become fluorescent upon re-exposure to the probe (data not shown). In contrast, when parental cells (containing the ~M transporter) were labeled with F-MTX and treated with PI-PLC, the fluorescence intensity remained constant (Fig. 8, panels C and D). It should also be noted that Figure 8 (compare panels A and C) provides visual evidence that the nM transporter in the JF subline is present at a much higher level than its/xM counterpart in the parental line. These results indicate that the L1210/.tM folate transporter is a 43 kDa integral membrane protein and does not contain asparagine-linked carbohydrate. In contrast, the /zM transporter from K562 cells (16), and possibly from various other human malignant cell lines as well, is heavily glycosylated and has a much higher molecular weight (ca. 80 kDa). Since the various murine and human/~M folate transporters have very similar kinetic properties (16), the rationale for this discrepancy in carbohydrate content and molecular weight is not apparent. The L1210 nM folate transporter is a 39 kDa glycosylated protein anchored exofacially on the plasma membrane by a GPI tail. In these properties it closely resembles its counterparts from human sources (e.g., KB cells (3), Caco cells (4) and placenta (5)) and from MA-104 monkey kidney cells (4). The nM transporters from KB cells, MA-104 cells and placenta have very similar amino acid sequences, and it is likely that the L1210 nM protein will not be too different.
12
J. FAN, et al. FUTURE DIRECTIONS
Based upon the results described above, it seems likely that biotin-SSMTX and biotin-SS-folate probes could be used to isolate folate transport proteins from other sources. Proteins responsible for folate efflux might also be identified by using the probes with membrane fragments, in which the inner membrane surface would be accessible. The general procedure might also be adapted to the identification and isolation of folate-dependent enzymes from cytoplasmic extracts. In both instances, however, functionality of the protein would not be demonstrable because of the covalent bond between the probe and the binding site. Future work should be directed toward development of a structure that would allow for reversible attachment of the probe. Another use of the biotin derivatives of MTX and folate is to visualize the folate transporters in individual cells. This has been accomplished by treating L1210 cells with these probes followed by a fluorescein conjugate of streptavidin. Results similar to those obtained with the fluorescein derivatives of MTX and folate (Fig. 8) are obtained (data not shown). Cross-labeling of the transporters (when both are present in the same cell) can be avoided by using: (a) activated biotin-MTX in the presence of a large excess of folate when only the/~M transporter is to be visualized; and (b) non-activated biotin-folate for exclusive labeling of the nM transporter. The use of fluorescein-streptavidin and rhodamine-streptavidin should allow color-specific labeling of the individual transporters to be obtained. One can envision many additional uses for these probes as visual markers of the folate transporters. These include: (1) To follow changes, if they occur, in the levels of these transporters during cell growth or cell cycle. (2) To assess the level of folate transporters in cells from cancer patients. This may be especially important in view of the report that the /zM folate transporter disappears during the DMSO-induced differentiation (and loss of malignancy) of murine erythroleukemia cells (17). It would be of interest to determine whether the /zM transporter appears during transformation of normal cells to tumor cells. (3) To investigate the mechanism for up-regulation of the nM folate transporter when cells are adapted to grow on low levels of folate. There are at least three possible explanations of this phenomenon: (i) selection of pre-existing cells with high levels of the transporter; (ii) loss of a folate-repressor protein complex that regulates expression of the transporter gene; or (iii) amplification of the transporter gene during growth perturbations caused by folate limitation. (4) To delineate the mechanism by which the nM transporter, which appears to be localized entirely outside of the membrane, translocates bound substrates through the membrane. Kamen and his colleagues (18) have proposed an endocytotic mechanism involving
FOLATE TRANSPORTSYSTEMSIN LI210CELLS
13
caveoli instead of coated pits without lysosomal participation, followed by transfer to the H,M transporter. SUMMARY Biotin derivatives of methotrexate (biotin-SS-MTX) and folate (biotin-SS-folate), in which the functional components are joined by a dissociable disulfide-containing spacer, have been synthesized, purified by DEAE-Trisacryl chromatography, and characterized by HPLC, elemental analysis and mass spectrometry. These compounds provide a convenient means for the single-step purification of the folate transporters from L1210 cells. Parental L1210 murine leukemia cells, which contain only the /.,M transporter (the reduced folate/MTX transport protein) were treated with the N-hydroxysulfosuccinimide ester of biotin-SS-MTX, and a detergent extract of the plasma membranes was exposed to streptavidin-agarose beads to adsorb the labeled protein. Dithiothreitol cleavage of the disulfide linkage released the transporter, which migrated as a well-defined component (43 kDa) on SDS-PAGE gels; no other proteins were present. An L1210 subline (JF), obtained by adapting cells to grow on nanomolar concentrations of folate, contains both the /zM transporter and the nM transporter (high-affinity folate binding protein). When these cells were treated with the N-hydroxysulfosuccimide ester of biotin-SS-folate and processed as described above, analysis on SDS-PAGE gels revealed the presence of two proteins, the /~M transporter (43 kDa) and the nM transporter (39 kDa). Both transporters were characterized with respect to amino acid content; blocked N-termini precluded Edman sequencing. Treatment of the nM transporter with peptide:N-glycosidase F produced a smaller component (32 kDa); the p.M transporter, conversely, was unchanged by this procedure. When the /zM transporter in parental L1210 cells was labeled with fluorescein-MTX and then treated with phosphoinositol-specific phospholipase C (PI-PLC), no change in fluorescence was detected. Alternatively, when the nM transporter in the JF subline-was labeled with fluorescein-folate and then treated with PI-PLC, complete loss of fluorescence was observed. These results indicate that the L1210/.tM transporter is a non-glycosylated, integral membrane protein, while its nM counterpart is heavily glycosylated and anchored exofacially to the membrane by a glycosylphosphatidylinositol component. ACKNOWLEDGEMENTS This investigation was supported by an Outstanding Investigator Grant (CA 39836) from the National Cancer Institute and a Grant (CH-31) from
14
J. FAN, et al.
the American Cancer Society. The authors are indebted to Ms Susan Burke for expert assistance in the preparation of the manuscript. REFERENCES 1. D. KESSEL, T. C. HALL, D. ROBERTS and I. WODINSKY, Uptake as a determinant of methotrexate response in mouse leukemias, Science 150, 752-754 (1965). 2. G . B . HENDERSON, E. M. ZEVELY and F. M. HUENNEKENS, Purification and properties of a membrane-associated folate-binding protein from Lactobacillus casei, J. Biol Chem. 252, 3760-3765 (1977). 3. P . C . ELWOOD, Molecular cloning and characterization of the human folate-binding protein cDNA from placenta and malignant tissue culture (KB) cells, J. Biol. Chem. 264, 14893-14901 (1989). 4. S. W. LACEY, J. M. SANDERS, K. G. ROTHBERT, R. G. W. ANDERSON and B. A. KAMEN, Complementary DNA for the folate binding protein correctly predicts anchoring to the membrane by glycosyl-phosphatidylinositol, J. Clin. Invest. 84, 715-720 (1989). 5. M. RATNAM, H. M A R Q U A R D T , J. L. DUHRING and J. H. FREISHEIM. Homologous membrane folate binding proteins in human placenta: cloning and sequence of a cDNA, Biochemistry 28, 8249-8254 (1989). 6. G. B. HENDERSON, Transport of folate compounds into cells, pp. 207-250 in Nutritional, Pharmacologic and Physiologic Aspects of Folates and Pterins. (R. L. BLAKLEY and M. V. WHITEHEAD, eds.) John Wiley and Sons, New York (1986). 7. G. B. HENDERSON, J. M. TSUJI and H. P. KUMAR, Mediated uptake of folate by a high-affinity binding protein in sublines of L1210 cells adapted to nanomolar concentrations of folate, J. Membrane Biol. 101, 247-258 (1988). 8. G. JANSEN, G. R. WESTERHOF, 1. KATHMANN, B. C. RADEMAKER, G. RIJKSEN and J. H. SCHORNAGEL, Identification of a membrane-associated folate-binding protein in human leukemic CCRF-CEM cells with transport-related methotrexate resistance, Cancer Res. 49, 2455-2459 (1989). 9. J. FAN, L. E. POPE, K. S. VITOLS and F. M. HUENNEKENS, Affinity labeling of folate transport proteins with the N-hydroxysuccinimide ester of the ~-isomer of fiuorescein methotrexate, Biochemistry 30, 4573-4580 (1991). 10. C.-H. YANG, F. M. SIROTNAK and L. S. MINES, Further studies on a novel class of genetic variants of the LI210 cell with increased folate analogue transport inward, J. Biol. Chem. 263, 9703-9709 (1988). 11. E.M. PRICE and J. H. FREISHEIM, Photoaffinity analogues of methotrexate as folate antagonist binding probes. 2. Transport studies, photoaffinity labeling, and identification of the membrane carrier protein for methotrexate for murine L1210 cells, Biochemistry 26, 4757-4763 (1987). 12. J. FAN, K. S. VITOLS and F. M. HUENNEKENS, Biotin derivatives of methotrexate and folate: synthesis and utilization for affinity purification of two membrane-associated folate transporters from L1210 cells, J. Biol. Chem. 266, 14862-14865 (1991). 13. J. FAN, L. E. POPE, K. S. VITOLS and F. M. HUENNEKENS, Visualization of folate transport proteins by covalent labeling with fluorescein methotrexate, Advan. Enzyme ReguL 30, 3-12 (1990). 14. U. KUEFNER, A. ESSWEIN, U. LOHRMANN, Y. MONTEJANO, K. S. VITOLS and F. M. HUENNEKENS, Occurrence and significance of diastereomers of methotrexate ct-peptides, Biochemistry 29, 10540-10545 (1990). 15. G. B. HENDERSON and E. M. ZEVELY, Affinity labeling of the 5methyltetrahydrofolate/methotrexate transport protein of L1210 cells by treatment with an N-hydroxysuccinimide ester of [3H]methotrexate, J. Biol. Chem. 2,59, 4558-4562 (1984). 16. L . H . MATHERLY, C. A. CZAJKOWSKI and S. M. ANGELES, Identification of a highly glycosylated methotrexate membrane carrier in K562 human erythroleukemia
FOLATE TRANSPORT SYSTEMS IN LI210 CELLS
15
cells up-regulated for tetrahydrofolate cofactor and methotrexate transport, Cancer Res. 51, 3420-3426 (1991). 17. R. E. CORIN, H. C. HASPEL and M. SONENBERG, Transport of the folate compound methotrexate decreases during differentiation of murine erythroleukemia cells, J. Biol. Chem. 259, 206-211 (1984). 18. B. A. KAMEN, A. K. SMITH and R. G. W. ANDERSON, The folate receptor works in tandem with a probenecid-sensitive carrier in MA104 cells in vitro, J. Clin. Invest. 87, 1442-1449 (1991).