Primary Structure and Cellular Distribution of Two

Val. 266, No. 9, Issue of March 25. pp. 5963-5972.1991 Printed in U.S.A.

THEJOURNALOF BIOLOGICAL CHEMISTRY (c)

1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Primary Structure and Cellular Distribution of Two Fatty Acid-binding Proteins in Adult Rat Kidneys* (Received for publication, August 13, 1990)

Hideki KimuraSQll, Shoji OdaniS, Shin-ichi NishiQ,Hirokazu SatoQ,Masaaki Arakawae, and Teruo OnoS From the Departments of $Biochemistry and §Medicine(II), Niigata University School of Medicine, Niigata 951, Japan

Fatty acid-binding proteins (FABPs) were purified from the kidneys of female and male rats and characterized by primary structure and histological distribution in the kidney. Two FABPs (14 and 15.5 kDa) were found in male rat kidney cytosol whereas only 14-kDa FABP could be recognized in female rat kidneys throughout the purification steps. The amino acid sequence of the 14kDa FABP was identical to that of rat heart FABP deduced from the cDNA sequence (Heuckeroth, R. O., Birkenmeier, E. H., Levin, M. S., and Gordon, J. I. (1987) J. Biol. Chem. 262,9709-9717). Structural analysis of the male-specific 15.5-kDa FABP identified this second FABP as a proteolytically modified form of 02,-globulin, an 18.7-kDa major urinary protein of adult male rats (Unterman, R. D., Lynch, K. R., Nakhasi, H. L., Dolan, K. P., Hamilton, J. W., Cohn,D. V., and Feigelson, P. (1981) Proc. Natl. Acad. Sci.U. S. A. 78,3478-3482) which shares a common ancestry with a numberof hydrophobic ligand-binding proteins such as serum retinol-binding proteins. Immunohistochemical investigation disclosed that heart-type FABP (14-kDa FABP) is localized in the cytoplasm of the epithelia of the distal tubules in both male and female rat kidneys whereas 15.5-kDaFABP immunostaining was observed predominantly in the endosomes or lysosomes of proximal tubules in male rat kidneys. These results suggest strongly the functional divergence of two FABPs in the rat kidney.

other cytosolic proteins such ascellular retinoid-binding proteins and peripheral nerve myelin P2 protein(6,11). Recently, two new proteins, bovine MDGI, a proposed growthinhibitor, and gastrotropin, a putative stimulator of gastric acid and pepsinogen secretion, were shown to belong to this family, extending the homology in an unexpected way (12, 13). The former is of particular interest because this growth inhibitor differs from bovine heart FABP by only a few residues (14). This finding appears to emphasize the necessityformuch more detailed examination of the chemical structure of the individualheartFABP-likeproteins, whichhave been so widely detected in many tissues by immunochemical (9, 15) or Northern blot analysis ( 7 ) . The kidney is one of the organs thatis actively involved in fatty acid metabolism. A considerable amount of fatty acid is taken up from blood and utilized at a high rate in thekidney (16,17) where FABPs are presumed to play a key role in their transport and metabolism. Earlier workers showed that rat kidney cytosol contains a binding protein for long chain fatty acids (1)and that the fatty acid-binding capacity of kidney cytosol increases in response to a hypolipidemic drug, clofibrate (18).Recently Lam et al. (19) reported the presence of two FABPs in male rat kidneys, one with a molecular mass of 14 kDa and the other witha molecular mass of 15.5 kDa. The 14-kDa proteinis closely related to heart-type FABP in its electrophoretic mobility and reactivity topolyclonal antiheartFABPantibodies.Ontheotherhand,the 15.5-kDa protein, designated “kidney FABP,” is clearly distinct from any of the tissue-specific FABPs in terms of molecular mass Fatty acid-binding proteins (FABPs),’ found in the cytosol and amino acidcomposition. Lam et al. also showed that differentiallyregulated, of various animal tissues, are low molecular mass (around 14 biosynthesis of the two FABPs is suggesting the existenceof functional differences between the kDa) proteins that are capable of binding long chain fatty we examined proteins (19). In a preliminary investigation (20) acids (1). They are thought to be involved in intracellular transport andmetabolism of fatty acids(2). Among them, rat the amino acid composition and amino-terminalsequence of liver, intestinal, and heart FABPs have been studied exten- kidneyFABPandfoundthatthisprotein is identical or sively (3-10). Structural data have revealed that they are closely similar to agu-globulin,which is a major male-specific homologous proteins and form a protein superfamily with protein in rat urine. However, detailed structures of the two FABPs in the kidney remain to be examined. Furthermore, *This work was supported in part by research grants from the Ministry of Education, Science, and Culture of Japan and the Naito since we were not able to detect 15.5-kDa kidney FABP in (20), another problem Foundation. The costs of publication of this article were defrayed in femalerat kidneyimmunologically part by the payment of page charges. This article must therefore be arose as to what kinds of FABPs are present in female rat hereby marked “advertisement” in accordance with18 U.S.C. Section kidneys. In particular, a survey for a possible female counter1734 solely to indicate this fact. V To whom correspondence should be addressed: Dept. of Biochem- part of the male 15.5-kDa FABP was very intriguing. In this paper, we describe the purification of FABPs from istry, Niigata University School of Medicine, 1-757 Asahimachi-dori, Niigata 951, Japan. Tel.: 025-223-616 (ext. 2262); Fax: 025-229-1808. female and male ratkidneys, the primary structureof female ’ The abbreviations usedare: FABP(s), fatty acid-binding pro- 14-kDa FABP, a detailed structural characterizationof male tein(s); ANS,8-anilino-1-naphthalenesulfonate; SDS, sodiumdodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; HPLC, high per- 15.5-kDa FABP, and their cellular and subcellular localizaformance liquid chromatography; ODs, octadecylsilane; mTALH, tions, which will be a clue to understanding functional differmedullary thick ascending limbof Henle; TFA, trifluoroaceticacid. ence between the two proteins in thekidney.

5963

5964

Two Fatty Acid-binding Proteins

Kidneys Adult Rat in

EXPERIMENTAL PROCEDURES~

Purification of a FABP from Female Rat Kidneys-Kidneys were perfused immediately with cold 0.25 M sucrose and excised from 25 female Wistar rats. The tissueswere minced with scissors, diluted to twice the volume with buffer A (10 mM Tris-HC1, pH 7.4, containing 10 mM KCI, 1 mM EDTA, and 1 mM dithiothreitol) and thenhomogenized with a Polytron homogenizer. This and all subsequent steps were performed at 4 "C and all buffer pH values given reflect the pH of the solution at 25 "C. The total homogenate was centrifuged at 3,000 X g for 15 min, and the resulting supernatant was centrifuged a t 105,000 X g for 2 h. The 105,000 X g supernatant containing approximately 360 mg of protein was concentratedto 40 ml by ultrafiltration (Amicon YM-5 membrane). The solution was separated from denatured proteins by centrifugation at 10,000 X g for 10 min and applied to a Sephadex G-75 column (4.9 X 95 cm)equilibrated with buffer A. The fractions containing high bindingactivityfor palmitate (in the 10-20-kDa range) were combined, dialyzed against buffer B (30 mM Tris-HC1, pH 8.5) and applied to a DEAE-cellulose column (1.8 X 10 cm) equilibrated with buffer B. The column was washed initially with the equilibrating buffer until unbound proteins were eluted completely. Then, retained proteins were eluted with a 100-ml linear gradientof 0-0.1 M NaCl in buffer B followed by a 100ml linear gradient of 0.1-0.3 M NaCl in buffer B. The fractions with high binding activity for palmitate were detected in the eluate with the former gradient. These fractionswere combined, dialyzed against 10 mM sodium acetate, pH 5.0, and chromatographed on a CMcellulose column (0.9 X 6 cm) equilibrated with the same buffer. The column was developed with a 100-ml linear gradient of0.01-0.3 M sodium acetate, pH 5.0. A homogeneous protein with high fatty acidbinding activity was eluted with this gradient. The purified protein was dialyzed against 30 mM Tris-HC1 buffer, pH 8.0, and used for further study.

600

800

1000

1200

z. I "

I

i I

0

100

1400

I

?

I P

200

p

RESULTS

Purification of FABPs from Cytosols of Male and Female Rat Kidneys Two species of FABP were purified from male rat kidneys t o homogeneity on SDS-PAGE by the procedure of Fujii et al. (21) (data not shown). One was a heart FABP-like protein (14-kDa FABP(m), "m" stands for male) and the other was the kidney FABP (15.5-kDa FABP) described by Fujii et al. (21) and by Lam et aL(l9).On the other hand, female kidney cytosol yielded only the 14-kDa FABP species (14-kDa FABP(f), "f' for female), which was prepared by a slight modification of the procedure of Fujii et al. (21). To monitor the purification of 14-kDa FABP(f), we measured the binding of palmitateand 8-anilino-1-naphthalenesulfonate(ANS). Fig. lA shows the binding of palmitateand ANS to the proteins of the 105,000 X g supernatant after gel filtration on Sephadex G-75. The profile of binding activity for ANS is quite similar to thatfor palmitate in this step. The firstlarge peak with fatty acid-binding activity corresponded to a molecular mass of about 60 kDa and maybe contaminating serum albumin. The second fatty acid-binding peak corresponded to molecular masses of 10-20 kDa and was considered to include a FABP. The FABP fractions were combined and then subjected to DEAE-cellulose column chromatography. Fig. 1B shows the profile of binding activity. A single peak with fatty acid-binding activity was identified in the fractions eluted with a linear gradient of 0-0.1 M NaC1. Since the peak still contained minor contaminants, it was purified further by cation-exchange chromatography on CM-cellulose (Fig. IC). During all steps of this purification, a peak corresponding to the 15.5-kDa FABP (kidney FABP) was not detected. ConPortjons of this paper (including part of "Experimental Procedures," Figs. 1S-6S, and Tables 1-6) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from WaverlyPress.

ELUTIONVOLUME

(ml)

FIG. 1. Purification of 14-kDa FABP(f) from female rat kidneys. A , elution profile of proteins from female rat kidney 105,000 X g supernatant from a Sephadex G-75 column (4.9 X 95 cm). The column was equilibrated and developed with 10 mM Tris-HC1, pH 7.4, containing 10 mM KCl, 1 mM EDTA, and 1mM dithiothreitol a t a flow rate of40 ml/h. B , combined fractions (bar) from A were chromatographed on a DEAE-cellulose column (1.8 X 10 cm) equilibrated with30 mM Tris-HC1, pH 8.5, and developed with a nonlinear gradient of 0-0.3 M NaCl in the equilibrating buffer (. . . . .) at a flow rate of 15 ml/h. C, combined fractions (bar) from B were dialyzed against 10 mM sodium acetate, pH 5.0, and chromatographed on a CM-cellulose column (0.9 X 6 cm) equilibrated with the same buffer and developed with a linear gradient of 10-300 mM sodium acetate, pH 5.0 (. . . . .), at a flow rate of 1 2 ml/h. In each purification step, the column eluate was monitored at 280 nm (@) and fractions were assayed for palmitate-binding (0)and ANS-binding(W) activities.

sequently, we purified only one FABP, namely 14-kDa FABP(f), from the cytosol of female rat kidneys. The results of a typical purification procedure are summarized in Table 1 (miniprint). The yield of 14-kDa FABP(f) was about lo%, which was comparable to thatobtained with human, rat, and pig heart FABPs (33). Since the profile of ANS-binding activity was very similar to thatof palmitate-binding activity in each purification step (Fig. 1, A X ) , we concluded that the ANS binding assay is applicable for monitoring the purification of 14-kDa FABP(f)

Two Fatty Acid-binding Proteins in Adult Rat Kidneys

5965

as well as Z protein (23), which is now identified as liver thechymotryptic digest,some of peaksrepresented pure FABP. peptides. Peak 7 in the tryptic digest and peaks 1 and 2 in Protein samplesfrom different purification stepswere sub- the chymotryptic digest were mixtures of two peptides, T7-1 jected to SDS-PAGE (Fig. 2). The purified 14-kDa FABP(f) and T7-2, C1-1 and C1-2, and C2-1 and C2-2, respectively, obtained after chromatography on CM-cellulose migrated as which were purifiedfurther by rechromatography on the same a single band with a molecular mass of 14 kDa, which was a column underslightlydifferent conditions. These purified little smaller than that reported for kidney FABP (19). When peptides were subjected to amino acid analysis (Tables 2 and theelectrophoretic mobilities of 14-kDa FABP(f) and rat 3, miniprint) and sequence analysis. On the basis of these heart FABPwere compared on SDS-PAGE, the two proteins results, all the tryptic and mostof the chymotryptic peptides migrated similarly (Fig. 2, lanes 4 and 5 ) . The amino acid could be unambiguously aligned with the cDNA-deduced secomposition of 14-kDa FABP(f) was also similar to that of quence of rat heart FABP (Fig. 3). This therefore demonheart FABP but clearly different from that of kidney FABP stratedthattheprimarystructure of 14-kDaFABP(f)is (datanotshown). Additionally, 14-kDaFABP(f) showed identical to thatdeduced from the cDNA for rat heart FABP cross-reactivity withpolyclonal anti-rat heart FABP IgG (Fig. (7, 34). Presently, 124 of 132 amino acids have been verified IS, miniprint). by directsequencing of isolatedpeptides.Compositional analysis of T13 indicated that this peptide covered the missing Primary Structure of 14-kDa FABP(f) 8 residues, but it was not susceptible to Edman degradation, We showed that 14-kDa FABP(f) wasclosely similar to rat presumably because of a blocked aminoterminus. Liquid heart FABP. Furthermore, to discuss the structural relation-secondary ion mass spectrometryof T13 identified a protonship between the two FABPs, we examinedtheprimary ated molecule of mle = 1036, which can be interpreted as an structure of 14-kDaFABP(f)andcompareditwiththat acetylatedform of thepeptide Ala-1 to Lys-9 (datanot predicted from the cDNA for rat heart FABP (7, 34). When shown). These results indicate that the amino terminus of 14the native 14-kDa FABP(f) was subjected to Edman degra- kDaFABP(f)isan acetylalanylresidue; i.e. theinitiator dation, no phenylthiohydantoin derivative was detected, in- methionyl residue had been removed. dicating that the amino terminuswas blocked. Thus 14-kDa Primary Structure of 15.5-kDa FABP KidneyFABP FABP(f) wasdigested with trypsin or chymotrypsin. The digests were separated by HPLC on an ODS reverse-phase In our preliminary study, we purified kidney FABP and column (Figs. 2 s a n d 3S, miniprint). In the tryptic digest, determined the amino acid composition and the amino-tereach peak except for peak 7 represented a pure peptide. In minal 20-residue sequence. On the basis of the similarity of amino acidcomposition and the identityof the amino-terminal sequence, we suggested the possible identity of kidney kDa FABP with a2,-globulin from which the amino-terminal 9residue segment had been removed (20). However, the molecular mass predicted from the sequence data for the processed a2,-globulin is 17.7 kDa, which is not consistent with the 21 molecular mass of 15.5 kDa determined by SDS-PAGE. This 14 lower value for kidney FABP was confirmed by HPLC gel permeation chromatography using a TSK-gel G3000SW col1 2 3 4 5 umn equilibrated with 6 M guanidine HCl (data not shown). FIG.2. Slab gel of electrophoresis of 14-kDa FABP(f). A Therefore, modification otherthanintheamino-terminal sample from each step of the purification procedure was electropho- region was suggested. T o obtain more structural information resed in a 10%polyacrylamide slab gel as described by Schaegger and region, 15.5-kDakidney von Jagow (24). Lane 1, 105,000 X g supernatant from female rat includingthecarboxyl-terminal kidney (20 pg). Lane 2, combined FABP-containing fractions from FABP was cleaved with cyanogenbromide. The CNBr-diSephadex G-75 gel filtration (5 pg). Lane 3, combined FABP-contain- gested protein was reduced with dithiothreitol and then aling fractions from chromatography on DEAE-cellulose (10 pg).Lane kylated with iodoacetamide. The resulting peptideswere sep4,14-kDa FABP(0 (5pg), the fraction from chromatography on CM- arated into nine peaks by HPLC on an ODS reverse-phase cellulose. Lane 5,rat heart FABP (5 pg). column (Fig. 4). Since peaks4 and 5 were collected together, themixture waspurified further by HPLConthesame I 10 20 30 Ae-A-D-A-F-V-G-T-W-K-L-V-D-S-K-N-F-D-D-D-Y-M-K-S-L-G-V-G-F-A-T-Rcolumn with a narrower gradient of acetonitrile '(Fig. 4S, .. .T13... . "T3-"-.TIO-TI-8 " c-c2-1--cl-lminiprint). All peaks (peaks 1-9) represented homogeneous 40 50 60 Q - V - A - S - M - T - K - P - T - T -- II - E - K - N - G - D - T - I -T-l -K-T-H-S-T-F-K-N-Tpeptides. These purified peptides were subjected to amino T9 -77-I---T5--, "C1-l" -c 10c7 acid analysis (Table 4, miniprint) and sequence analysis. It 70 80 90 E - I - S - F - Q - L - G - V - Q - F - D - E - V - T - A - D - D - R - K - V - K - S - V - V - T - L - D - G - G - Kwas noted that CNBr fragments of the same amino acid 72-T-" - C I 4 - - " " - ~ c IO" compositions were eluted from the reverse-phase column as IO0 110 120 L - V - H - V - Q - K - W - D - G - Q - E - T - T - L - T - R - E - L - S - D - G - K -- LL--TI- L - T - H - G pairs ( i e . CB1 and CB2, CB6 and CB7, and CB8 and CB9). -T6--T"T2---TI2This appears to be due to partial opening of the terminal "cI-2--"-" c 3 -c 2-2c-"21 -c4homoserine lactone rings, which should decrease the hydro130 D-V-V-S-T-R-T-Y-E-K-E-A phobicity of the peptides,leading to the behavior on the "T12" T I -T 1column. c 2-2"-c -4 * On the basis of these results, we assigned all of the CNBr FIG.3. Alignment of the tryptic and chymotryptic peptides peptides except for CB4 and CB5 with residues 10-117 of the from 14-kDa FABP(f) withthe amino acid sequencededuced sequence predicted from the cDNA for mature ap,-globulin from the cDNA for rat heart FABP (7, 34). Standard single- (35) (Fig. 5). letter abbreviations are used, and Ac represents N-acetyl. Peptides Amino-terminal 10-residue sequence analysis of peptides identified by automatic Edman degradation are indicated by underscored double arrows. The peptide T13 identified by amino acid CB4 and CB5 yielded a single unambiguous sequence corresponding to residues 118-127 of the predicted sequence. The composition and mass spectrometry is indicated by a dotted line. 4 "

-

--

"

-

"

P

PIroteins in Adult Rat Kidneys

Acid-binding Two Fatty

5966

glutamine and alanine and between alanine and arginine is uncommon and unprecedented, we judged CB4 and CB5 to be the carboxyl-terminal peptides generated from purified kidney FABP. Therefore, it is most likely that kidney FABP (15.5-kDa FABP) undergoes proteolytic processing in the carboxyl-terminal region as well as in the amino-terminal region and loses the carboxyl-terminal 2- or 3-residue segment of the predicted sequence of a2,-globulin. 0

10

30 40 50 60 ELUTION TIME (min)

70

20

Location of the Disulfide Bond of Kidney FABP

Reaction of kidney FABP with iodoacetamide under denaturing and nonreducing conditions yielded approximately 1 nmol of cyanogen bromide-digested kidney FABP was reduced and mol of cysteine/mol of protein whereas S-carbamoylmethyalkylated by S-carbamoylmethylation with iodoacetamide. The re- lation under denaturing and reducing conditions gave 3 alsulting peptideswere separated on a Toyo Soda ODS column (0.46 x kylated cysteine residues. The protein migrated as a single 25 cm)withthe following solvent gradient:solvent A, 1% (v/v) band on reducing or nonreducing SDS-PAGE (datanot acetonitrile in 0.1% (v/v) trifluoroacetic acid solvent B, 75% (v/v) acetonitrile in 0.1% (v/v) trifluoroacetic acid; 0 min, 0% B; 9.5 min, shown). Theseresults suggested that the purified kidney FABP has one disulfide bond and one free cysteine. 0% B; 32 min, 30% B; 79.5 min, 58% B; the flow rate, 1 ml/min. To determine the location of the disulfide bond, a thermolLarge peaks before CB1 are due to reagents for reduction and Scarbamoylmethylation. ysin digest of native kidney FABP (25 nmol) was separated by HPLC with a linear gradient of solvent B from 0 to 80% 1 10 20 30 in 60 min (data not shown). Amino acid analysis after perE-E-A-S-S-T-R-G-N-L-D-V-A-K-L-N-G-D-W-F-S-I-V-V-A-S-N~K-R-ECB6 formic acid oxidation of each fraction revealed that fraction CBT Th8 (yield, 58.9%) contained acystine peptide. Since fraction 40 50 60 K - I - E - E - N - G - S - M - R - V ~ F - M - Q - H - I - D - Y - L - E - N - S - L - G - F - K - F - R - I - K - E - Th8 included several contaminating peptides, it was further t C B 1-C C B 8 - p purified by HPLC on the same column with a shallower t CB 2 + -C B Q gradient, and peptide Th8-4 was isolated as a pure cystine IO 80 90 N-G-E-C-R-E-L-Y-L-V-A-Y-K-T-P-E-D-G-E-Y-F-V-E-Y-D-G-G-N~T-Fpeptide (Fig 6S, A , miniprint). The amino acid composition CBB of the peptide showed that it was a pure 16-residue peptide 100 110 120 T-I-L-K-T-D-Y-D-R-Y-V-M-F-H-L-I"F-K-N-F-K-N-~-E-T-F-Q-L-M-V-L-Y(Table 6, miniprint). Peptide Th8-4 was oxidized with perCB3 -, formic acid and then subjected to HPLC under the same t K 4t " conditions as used for further purification of Th8. Two major t K 4peptides, Th8-4aandTh8-4b, were obtained (Fig. 6S, B, 130 140 150 G-R-T-K-D-L-S-S-D-I-K-E-K-F-A-K-L-C-E-A-H-G-I-T-R-D-N-I-I-Dminiprint), corresponding to residues 151-157 and 58-66, C B 4 -K4-t-K5--"-,+Kl+tK2+4 K6 respectively (Table 6, miniprint). The results indicated the ". structure of Th8-4 asfollows: -K 4 " K5 +K I + t K 2 + K6 FIG. 4. Reverse-phase HPLC separation of peptides obtained from cyanogen bromide cleavage of kidney FABP. 20

. "

7

"

7

7

"

~

"

"

7

"

"

7

~

t " " " 7 -

. b

7 - 7 7

4

" 7 7 " " "

-C

"LI

.t7

77""

7

7 7 7

" " " 7

" " " 7

4

4

160 L-T-K-T-D-R-C-L-Q-A-R-G " C B 4 -

CK B 5 "--+-K3b-

__f

-c

3

8-

FIG. 5. Alignment of the CNBr-derived peptides from kidney FABP with the amino acid sequence of mature az.-globulin reported by Unterman et al. (35).Peptides identified by amino acid composition are indicated by underscored double arrows. Residues identified by automatic Edman degradation are indicated by half-arrows.

amino acid compositions of the two peptides wereclosely similar to that of the carboxyl-terminal peptide, residues 118162 of a2,-globulin. However, peptides CB4 and CB5 were different from the expected carboxyl-terminal peptide in lacking 3 residues (1residue each of alanine, arginine, and glycine) and 2 residues (one arginine and one glycine), respectively (Table 4, miniprint). To confirm these minor differences, peptides CB4 and CB5were further analyzed structurally. Because of the limited amount of materialin Fig.4, the mixture of CB4 and CB5 was digested directly with lysyl endopeptidase, and the resultant peptides were separated by HPLC as shown in Fig. 5 s (miniprint). Based on the amino acid compositions of the isolated peptides (Table 5, miniprint), K3a and K3b were found to be the expected carboxylterminal peptides, residues 154-159 from CB4 and residues 154-160 from CB5, respectively. Other peptides (Kl, 2, K46) entirely covered the remaining regions of CB4 and CB5 (Fig. 5). From these results, the two peptides, CB4 and CB5, could be aligned precisely with residues 118-159 and 118-160, respectively. Since CNBr cleavage of peptide bonds between

Ile-Lys-Glu-Asp-Gly-Glu-Cys-Arg-Glu

(Th8-4b)

I Leu-Thr-Lys-Thr-Asp-Arg-Cys

(Th8-4a)

Thus, the disulfide bond of kidney FABP was deduced to be located between Cys-64 and Cys-157. Furthermore, 1 free cysteine was deduced to be at position 138 although we failed to detect a peptide containing free cysteine.

Immunohistochemical Localizations of 14-kDa FABP and Kidney FABP 14-kDa FABP-We used rabbit polyclonal IgG raised against heart FABPfor immunohistochemically localizing 14kDa FABP in rat kidneys utilizing the identity of 14-kDa FABP with rat heart FABP. Light microscopic immunohistochemical studies revealed the cell-specific distribution of heart FABP. In the medulla, the medullary thick ascending limb of Henle (mTALH), the initial portion of the distal tubule, was definitely stained by this polyclonal antibody (Fig. 6A). In the cortex, the immunoreactivity was predominantly observed in the epithelia of cortical distal tubules whereas the glomeruli and proximal tubules were scarcely stained (Fig. 6B). No difference between male and female specimens was observed inthe distribution of heart FABP. Immunogold labeling disclosed that the gold particles were positive in the cytoplasm of the epithelia of mTALH and cortical distal tubules (data not shown). Kidney FABP-In light microscopic immunohistochemical studies of male specimens, intensive staining was observed in

Two Fatty Acid-binding Proteins in Adult Rat Kidneys :

5967

,",",,~q?m?"T~-=mq .

.

Ft

'?

,'

"1

.

,

.

FIG. 7. Immunoelectron micrograph of the proximal tubule in male rat kidney stained with immuneIgG against kidney FABP. The immunoreactivity for kidney FAHI' is predominantly observed in the endosomes or lysosomes (a typical image is indicated by an arrowhead) of the epithelium of the proximal tubule as compared with the cytoplasm. The luminal side of the tubular epithelium is at the top of the figure. M , mitochondria. Magnification, X 10,000.

protein (5% of cytosolic proteins (19))may be of quantitative significance in fatty acid-binding activity despite its somewhat lower affinity for long chain fatty acids. The 14-kDa FABP(f) constitutes about 0.3% of the cytosolic proteins in female rat kidneys, which is 1 order of magnitude lower than F I ~ : 6. . Immunolightmicrographs of maleratkidney muscle (9,15). Theconcenstained with immune IgG against rat heart FABP. A , in the that of heart FABP in rat cardiac medulla ( m ) ,immunostaining lor heart FAHP is found diffusely in tration of 14-kDa FABP(m) in male rat kidneys was similar the epithelia of mTALH, which are major components in the medulla to thatin female rat kidney. whereas in the cortex (c) less staining is observed. R, in the cortex, ImmunochemicalstudiesandNorthern blot analyses the staining is observed in the distal tubules (asterisks). In contrast, showed that a heart FABP-like protein expressed is in various no or scarce immunoreactivity isseen in the glomeruli (g) and proximal tubules. In A and B, the sections are counterstained with rat tissues such asslow twitch skeletal muscle, kidney, testis, and brain (7,9, 15,34). We also detected andisolated a heart hematoxylin for nuclei. Magnifications: A, X 6; R,X 200. FABP-like protein from rat gastric mucosa (29). As for rat a cellular RNA the epithelia of the proximal tubules and the urinary lumina kidneys, it was shown that the tissue contains in thecortex. In the medulla, weak staining was found in the that can be protected completely from S1 nuclease digestion coding region for a carboxyl-terminal urinary lumina. In contrast, no positive staining was found in by a probe including the 68-residue segment of rat heart FABP (34). Recently, Lam et female rat kidney (data not shown). Immunogold labeling was performed to determine the subcellular localization of kidney al. (19) and we (20) have isolated a 14-kDa FABP similar to heart FABP from male rat kidneys. Strictly speaking, howFABP. Gold labelingwasobserved predominantlyinthe endosomes or lysosomes of the proximal tubular epithelium ever, these reports merely indicated thepresence of a protein in rat kidneys. In the present as compared with the cytoplasm (Fig. 7). Gold particles were closely similar to rat heart FABP study we present the first convincing evidence, from a detailed also present at thebrush-border membranes of the proximal of 14-kDa tubules (data not shown). On the other hand, no labeling was structural investigation, that the primary structure FABP(f) is identical to that of rat heart FABP (7,34). Some found in the epitheliaof the distal tubules, including mTALH discrepancies were evident between the primary structures of (data not shown). rat heart FABP determined by cDNA (7, 34) and protein (6, 8) sequencing, but we reexamined the primary structure of DISCUSSION rat heart FABP by protein sequencing and found that the We detected and purified only one FABP, namely 14-kDa positions inconflict were occupied by the aminoacid residues FABP(f),from the cytosol of female rat kidneysby monitoring in the cDNA-derived sequence (36). Thesequence of 14-kDa FABP(f) determined here is also identical to that deduced the palmiticacid- andANS-binding activities. Failureto detect any FABP other than 14-kDa FABP(f) in the cytosol from the nucleotidesequence, indicating that heart FABP with the same primary structure is actually present in the of female rat kidneys in the purification not only supports our previous observation that kidney FABP could not be kidney as themajor FABP. Tryptic and chymotryptic peptide maps of 14-kDa detected immunologically infemale rat kidneycytosol but also implies that female rat kidneys contain 14-kDa FABP(f) FABP(m) frommale rat kidneys were completely identical to as the only (or at least the predominant) species of FABP. that of female rat kidney 14-kDa FABP (figures not given). On the other hand, male rat kidneys contain two FABPs, with Therefore, it isprobable that heart FABP iscommon to both 15.5-kDa FABP (kidney FABP) being the major FABP (19, male and female rat kidneys. The structural analysisof kidney FABP (15.5-kDa FABP) 20). It seems that the rather high concentration of 15.5-kDa

5968

T w o Fatty Acid-binding Proteins

has proven that theprotein is identical with az,-globulin from which the carboxyl-terminal 2 or 3 residues as well as the amino-terminal 9residues have been removed. Other parts of the molecule are in agreement with the cDNA-deduced sequence for a2,-globulin (35). The molecular mass of this modified protein is calculated to be 17.1 kDa, which is not consistent with the observed value of 15.5 kDa. The possibility of further shortening in the carboxyl-terminal regionwas excluded by the isolation of the two small peptides (K3a and K3b, Fig. 5 s and Table 5, miniprint) from the carboxyl end of the molecule. Furthermore,all the inner parts of the molecule had been recovered as CNBr fragments, and their terminal sequences were confirmed (Fig. 5). Consequently, the discrepancy between the observed and calculated molecular masses remains unclear. a2,-Globulin is a representative of the recently proposed ap,-globulin protein superfamily, which consists of about a dozen proteins of molecular mass approximately 20 kDa (37). Each member of this superfamily is thought to be a hydrophobic molecule transporter, e.g. serum retinol-binding protein for retinol, apolipoprotein D for cholesterol. The known crystal structures of four members (@-lactoglobulin,serum retinol-binding protein, and two insect bilin-binding proteins) revealed that they uniquely share a @-barrel formed by two sets of four antiparallel 8-strands and an a-helix consisting of about 10 residues near the carboxyl terminus (38, 39). When the secondary structure of kidney FABP was evaluated by the Chou-Fasman method from the primary structure confirmed here, the analysis predicted that kidney FABP contains six regions of a-helix (residues 10-14, 29-32, 48-49, 56-58,114-119, and 131-141) and nine regions of /?-structure (residues 20-24, 39-47, 54-55,66-73,81-82,91-93, 101-107, 120-123, and 149-152). The predicted secondary structure of kidney FABP resembled that of &lactoglobulin in thenumber and position of @-structure and in having an a-helix composed of about 10 residues in the carboxyl-terminal region. Kidney FABP was also shown to have the disulfide bond between Cys-64 and Cys-157, which is conserved in several members of the a2,-globulin protein superfamily. @-Lactoglobulincontains two disulfide bonds, at 106-119 and 66-160 (40). The latter corresponds to the disulfide bond in kidney FABP. Serum retinol-binding protein has threedisulfide bonds (41). The disulfide linkage at 70-174 is equivalent to thatof kidney FABP. From these findings, this conserved disulfide bridge appears to be important in maintaining the three-dimensional structure for binding hydrophobic ligands. Immunohistochemical studies using anti-heart FABP IgG showed the distribution of heart FABP within renal tubular cells. Unexpectedly, heartFABP was scarcely present in proximal tubules, for we guessed that the proximal tubules would contain FABPs in large amount judging from the abundance of mitochondria in which &oxidation of fatty acids is carried out.Heart-typeFABP was abundantinthe mTALH. Recently, Chamberlin and Mandel (42) reported, using a suspension of distal tubules of rabbit kidneys, that the most effective endogenous substrates for ATP production inthe mTALH are probably fatty acids. Therefore, it is reasonable that heart-type FABP is diffusely present in the mTALH. Heart FABP in mTALH may be involved in fatty acid metabolism coupled with energy production for the tubules. Furthermore, the binding protein may also function as a protector for the epithelium of mTALH when kidneys suffer from hypoxic injury. It has been suggested that mTALH is particularly susceptible to hypoxic injury because of the perilously low oxygen supply to that segment (43). It is also known that a cascade of events including breakdown of mem-

Adult in

Kidneys Rat

brane phospholipids with release of free fatty acids such as arachidonic acids begins during hypoxia and that the products of lipid breakdown such as free fatty acids are cytotoxic (44). Moreover, Weinberg and Humes (45) reported that trapping of free fatty acids with fatty acid-free albumin could diminish the severity of mitochondrial membrane injury (45). These findings imply that heart FABPmay take a partin protecting mTALH from hypoxic injury. Immunoelectron microscopy of kidney FABP hasdisclosed the characteristic subcellular distribution within the proximal tubular cells in male rat kidneys. Kidney FABP is predominantly localized in the endosomes or lysosomes and is also associated with the brush-border membranes. This subcellular localization strongly suggests that in the proximal tubules kidney FAQP is reabsorbed by endocytosis with lysosomal degradation, a process that is comparable to that for low molecular mass proteins such as insulin (46). a2,-Globulin, which is a precursor of 15.5-kDa kidney FABP, is a secretory 18.7-kDa protein synthesized in the liver (35, 47), and no mRNA for the proteincan be detected in the kidney by Northern blot analysis (48). A recent report of Neuhaus (49) suggested that az,-globulin in urine is reabsorbed by a mechanism including the interaction of cationic groups within the molecule with anionic sites on the brush-border membranes. Taking these into account, it is almost certain that kidney FABP undergoes endocytotic uptake in the proximal tubules. Although kidney FABP was localized predominantly in endosomes or lysosomes, we and earlier workers purified the protein from the cytosol of male rat kidneys in large amount (19-21). Hence, kidney FABP thus purified may have originated mainly from endosomes that can be ruptured easily during preparationof kidney cytosol because of their fragility, but not from lysosomes in which proteins are generally degraded to small peptides or amino acids. During the time that az,-globulin is taken up from the urinary lumen into endosomes, it may be proteolytically processed in the amino- and carboxyl-terminal regions andthen converted to kidney FABP by the protease associated with the brush-border membrane. Lam et al. (19) noted “a slightly larger (about 16-kDa) and less abundant protein” immunologically cross-reactive to anti-kidneyFABP antibodies in male rat kidney cytosol, which may suggest sequential proteolytic processing in the kidney. However, since processing may also be attributed to an artifact of the isolation procedure, the purification of kidney FABP in the presence of several protease inhibitors will be necessary to determine whether the processing is an artifact or a physiological phenomenon. The apparent absence of 15.5-kDa kidney FABP or its equivalent in female rat kidneys seems to cast doubt on the proposed physiological function in lipid metabolism in the kidney suggested by developmentally-regulated(19) or hypertension-responsive (19, 50) alteration of expression. Alternatively, complex regulation by sex hormones of rat w,-globulin gene expression (51) and the presence of a female-specific hamster protein in the vaginal discharge, aphrodisin, which is a homolog of az,-globulin and acts as an aphrodisiac pheromone (52), may suggest a similar function as a pheromone transporter or pheromone itself for 15.5-kDa kidneyFABP. Our present resultsprovide conclusive evidence that kidney 14-kDa FABP is virtually the sole FABP in female rat kidney cytosol, is identical with heart FABP, and is also present in male rat kidney. Furthermore, we show that themale-specific 15.5-kDa kidney FABP is a processed form of a2,-globulin. Immunohistochemical examinations revealed differential celland organelle-specific localization of the two FABPs in male rat kidneys. However, the precise metabolic functions of the

Two Fatty Acid-binding Proteins two FABPs remain unclear, and we are continuing to investigate their physiological functions in thekidney. Acknowledgments-We are indebted to Drs. Satoshi Fujii and Hideaki Kawaguchi of Hokkaido University for supplying the antikidney FABP IgG fractions and to Professor Yasuo Takahashi of Niigata University for permitting us to use his laboratory facilities. We also thank Ken-ichi Sizukuishi of HitachiTechno-Research Center (Ibaragi, Japan) for liquid secondary ion mass spectrometry. We gratefully acknowledge the helpful discussion of Dr. Masahiro Hitomi (Niiaata - Universitv).

REFERENCES 1. Ockner, R. K., Manning, J . A., Poppenhausen, R. B., and Ho, W. K. L. (1972) Science 177, 56-58 2. Glatz, J. F. C., and Veerkamp, J. H. (1985) Znt. J. Biochem. 17, 13-22 3. Takahashi, K., Odani, S., and Ono, T. (1983) Eur. J. Biochem. 36, 589-601 4. Sweetser, D. A,, Birkenmeier, E. H., Klisak, I. J., Zollman, S., Sparkes, R. S., Mohandas, T., Lusis, A. J., and Gordon, J. I. (1987) J . Biol. Chem. 262,16060-16071 5. AlpersD.H., Strauss, A. W., Ockner, R. K., Bass, N. M., and Gordon, J. I. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,313-317 6. Sacchettini, J . C., Said, B., Schulz, H., and Gordon, J. I. (1986) J. Biol. Chem. 261,8218-8223 7. Heuckeroth, R. O., Birkenmeier, E. H., Levin, M. S., and Gordon, J. I. (1987) J. Biol. Chem. 262, 9709-9717 8. Gibson, B. W., Yu, Z., Aberth, W., Burlingame, A. L., and Bass, N. M. (1988) J. Biol. Chem. 263, 4182-4185 9. Bass, N. M., and Manning, J. A. (1986) Biochem. Biophys. Res. Commun. 137,929-935 10. Said, B., and Schulz, H. (1984) J. Bid. Chem. 259, 1155-1159 11. Takahashi, K., Odani, S., and Ono, T. (1982) Biochem. Biophys. Res. Commun. 106, 1099-1105 12. Bohmer, F.-D., Kraft, R., Otto, A., Wernstedt, C., Hellmann, U., Kurtz, A,, Muller, T., Rohde, K., Etzold, G., Lehmann, W., Langen, P., Heldin, C.-H., and Grosse, R. (1987) J. Biol. Chem. 262, 15137-15143 13. Walz, D. A., Wider, M. D., Snow, J. W., Dass, C., and Desiderio, D. M. (1988) J . Biol. Chem. 263, 14189-14195 14. Billich, S.,Wissel, T.,Kratzin, H., Hahn, U., Hagenhoff, B., Lezius, A. G., and Spener, F. (1988) Eur. J. Biochem. 175, 549-556 15. Crisman, T. S.,Claffey, K.P., Saouaf, R., Hanspal, J., and Brecher, P. (1987) J. Mol. Cell. Cardiol. 19, 423-431 16. Hohenleitner, F. J., andSpitzer, J . J. (1961) Am. J.Physiol. 200, 1095-1098 17. Barac-Nieto, M., and Cohen, J. J. (1971) Nephron 8,488-499 18. Paulussen, R. J. A., Jansen, G. P. M., and Veerkamp, J . H. (1986) Biochim. Biophys. Acta 877, 342-349 19. Lam, K. T., Borkan, S., Claffey, K. P., Schwarz, J . H., Chobanian, A. V., and Brecher, P. (1988) J. Biol. Chem. 263, 15762-15768 20. Kimura, H., Odani, S., Suzuki, J., Arakawa, M., and Ono, T. (1989) FEBS Lett. 246, 101-104 21. Fujii, S.,Kawaguchi, H., and Yasuda, H. (1987) Arch. Biochem. Biophys. 2 5 4 , 552-558 22. Glatz, J. F. C., and Veerkamp, J. H. (1983) Anal. Bichem. 132, 89-95

in AdultKidneys Rat

5969

23. Sugiyama, Y., Iga, T., Awazu, S., and Hanano, M. (1980) Biochem. Pharmacol. 29, 2063-2069 24. Schaegger, H., and von Jagow, G. (1987) Anal. Biochem. 166, 368-379 25. Hirs, C. H. W. (1967) Methods Enzymol. 11, 197-203 26. Glatz, J . F. C.,Baenvaldt, C. C. F., Veerkamp, J . H., and Kempen, H. J . M. (1984) J . Biol. Chem. 259,4295-4300 27. Matsubara, H., and Sasaki, R. M. (1969) Biochem. Biophys. Res. Commun. 35,175-181 28. Ouchterlony, 0. (1958) Prog. Allergy 5 , 1-78 29. Kanda, T., Iseki, S., Hitomi, M., Kimura, H., Odani, S., Kondo, H., Matsubara, Y., Muto, T., and Ono, T. (1989) Eur. J . Biochem. 185, 27-33 30. Hsu, S.-M., Raine, L., and Fanger, H. (1981) J . Histochem. Cytochem. 29,577-580 31. Uchida, T., and Endo, T. (1988) J. Histochem. Cytochem. 36, 693-696 32. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J . (1951) J. Biol. Chem. 193,265-275 33. Paulussen, R. J. A., van der Logt, C. P. E., and Veerkamp, J. H. (1988) Arch. Biochem. Biophys. 2 6 4 , 533-545 34. Claffey, K. P., Herrera, V. L., Brecher, P., and Ruiz-Opazo, N. (1987) Biochemistry 26, 7900-7904 35. Unterman, R. D., Lynch, K.R., Nakhasi, H. L., Dolan, K. P., Hamilton, J. W., Cohn, D.V., and Feigelson, P. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 3478-3482 36. Kimura, H., Hitomi, M., Odani, S., Koide, T., Arakawa, M., and Ono, T. (1989) Biochem. J. 260, 303-306 37. Pevsner, J., Reed, R. R., Feinstein, P. G . , and Snyder, S. H. (1988) Science 2 4 1 , 336-339 38. Holden, H. M., Rypniewski, W. R., Law, J. H., and Rayment, I. (1987) EMBO J . 6, 1565-1570 39. Huber, R., Schneider, M., Mayr, I., Muller, R., Deutzmann, R., Suter, F., Zuber, H., Falk, H., and Kayser, H. (1987) J . Mol. Biol. 198,499-513 40. Papiz, M. Z., Sawyer, L., Eliopoulos, E. E., North, A. C. T., Findlay, J. B. C., Sivaprasadarao, R., Jones, T. A,, Newcomer, M. E., and Kraulis, P. J. (1986) Nature 324, 383-385 41. Newcomer, M. E., Jones, T. A., Aqvist, J., Sundelin, J., Eriksson, U., Rask, L., and Peterson, P. A. (1984) EMBO J . 3, 14511454 42. Chamberlin, M. E., and Mandel, L. J. (1986) Am. J. Physiol. 25 1,F758-F763 43. Brezis, M., Rosen, S., Silva, P., and Epstein, F. H. (1984) Kidney Znt. 26, 375-383 44. Matthys, E., Patel, Y., Kreisberg, J., Stewart, J. H., and Venkatachalam, M. (1984) Kidney Znt. 26, 153-161 45. Weinberg, J. M., and Humes, H. D. (1985) Am. J. Physiol. 248, F876-F889 46. Hellfrizsch, M., Christensen, E. I., and Sonne, 0. (1986) Kidney Znt. 29, 983-988 47. Roy, A. K., and Neuhaus, 0. W. (1966) Biochim. Biophys. Acta 127,82-87 48. MacInnes, J . I., Nozik, E. S., and Kurtz, D. T. (1986) Mol. Cell. Biol. 6, 3563-3567 49. Neuhaus, 0. W. (1986) Proc. SOC.Exp. Biol. Med. 182, 531-539 50. Fujii, S.,Kawaguchi, H., Okamoto, H., Togashi, H., Saito, H., and Yasuda, H. (1988) J. Hypertension 6,671-675 51. Kurtz, D. T., Sipple, A. E., Ansah-Yiadom, R., and Feigelson, P. (1976) J . Biol. Chem. 251, 3594-3598 52. Henzel, W. J., Rodriguez, H., Singer, A. G., Stults, J., Macrides, F., Agosta,W. C., and Niall, H. (1988) J. Biol. Chem. 263, 16682-16687 ~" ~

~~

5970

Two Fatty Acid-binding Proteinsin Adult Rat Kidneys

105,0008 s u w r n a t a n t Seilhadex

0.5'

360b

G ~ 7 5

34.Zb

D E A E cellulose

4.2d

c

l.ld

"CellYloSe

100 41.8 17.0 9.7

2.2 7 3 15.9

1.0 4.4

14.6 31.8

'Measured r l t h I l - C " l ~ a l m l t i m ~ i d a1 37'C. bDetermlnedbythemethod O f Lowry e t . a i 1321. CDetermLned after dealbumlnation. dDetermhned from w s n t l t a t i v e arn~no acld analys~s.

hnlm

TZ

73

TI

T5

T6

T7-1 T 7 ~ 2

T8

TT9 I 3T I 2T I 1T I 0

achd Number

him acid

C I ~ I GI-2

cz-1 c 4 c 3c 2 - 2

of

residues ~ e molecule r of mrrtide'

c7

C8

CIO

Number of residues ~ e rmoleciile o f w p t l d e '

C12

CI4

Two Fatty Acid-binding Proteins inKidneys Adult Rat

Total

4

4

Parlllon

39-42

39-42

YleldlX)

45

15

42

28

43

53

60

29

118-160 10-38

118-159

103.117

45

29 10-38 45

60

5971

60

43-102 43-102 45

38

lT t

1

Total

27

3

7

6

7

I7

0 0

20

40

ELUTION TIME (min)

CBS

o.,oc

60

5972

Two Fatty Acid-binding Proteins in Adult Rat Kidneys