Tubulointerstitial nephritis (TIN) antigen has been re- cently identified as a novel basement membrane macro- molecule. It consists of a single chain of 58 kDa ...
THEJOURNAL OF B I O ~ I C A CHEMISTRY L. 0 1994 by T h e American Society for Biochemistry and Molecular Biology, Inc.
Vol. 269, No. 3, Issue of January 21, pp. 1654-1659, 1994 Printed in U.S.A.
Tubulointerstitial Nephritis Antigen Interacts with Laminin and Type IV Collagen and Promotes Cell Adhesion* (Received for publication, June 16, 1993, and in revised form, August 31, 1993)
Theodosia A. Kalfa, Jennifer D. “hull, Ralph J. Butkowski, and AristidisS . Charonis+ From the Department of Laboratory Medicine and Pathology, University of Minnesota Medical School, Minneapolis, Minnesota 55455 -
Tubulointerstitial nephritis (TIN) antigen has rebeenextent in the ileum. It is also present in small amounts inthe cently identifiedas a novel basement membrane macro-corneal and the epidermal basement membrane (2). The unique molecule. It consists of a single chain of 58 kDaand pattern of distribution of TIN-ag suggests that it may contribexhibits a restricted distribution. The interaction beute to a specific structural conformation of these basement tween TIN antigen and lamininor type IV collagen has membranes, possibly related to specialized function(s) of the beenstudiedusingsolid-phasebindingassaysand overlying cells. Therefore, it is of interest to examinethe interfound to be for both macromolecules specific, saturable, actions between thisnovel macromolecule and other basement and with an affinity in the low micromolar range. In membrane components in order to understand how the pressimilar assays,TIN antigen did not interact with hepa- ence of TIN-ag could modify the basement membrane ultrarin. In turbidimetry assays, it was found that the pres- structure. Furthermore, it is important to initiate studies on ence of TIN antigen did not affect the polymerization of the interactions between TIN-ag and various cell types in order a concentration-dependentintype IV collagen but had hibitory effect on laminin polymerization and on pre- to eventually correlateits presence with specific cellular funcformed laminin polymers.TIN antigen was able to pro- tions. mote adhesion of epithelial cells derived from kidney EXPERIMENTALPROCEDURES tubules and of endothelial cells derivedfrom aorta. The Isolation Procedures data suggest thatTIN antigen may be a macromolecule of importance both for basement membrane ultrastruc- TIN-ag was extracted and purified from rabbit kidney cortex baseture and cellular adhesion. ment membrane (BM) with a modification of the non-denaturating method previously described (2). To isolate BM, frozen rabbit kidney cortices (Pel-Freez Biologicals, Rogers, A R ) were thawed at 4 “Cby immersion in protease inhibitor buffer (10 m~ Tris-HC1, 0.15 M NaCl, Tubulointerstitial nephritis antigen (TIN-ag)’ has recently 1.0 m~ EDTA, 2.0 m~ e-aminocaproic acid,2.0 m~ N-ethylmaleimide, been described as a 58-kDa basement membrane component pH 7.4). Cortices were homogenized with a Polytron tissue disrupter recognized by human autoantibodies in certain formsof tubu- and thenpassed through a 35-mesh sieve.The filtrate was concentrated lointerstitial nephritis (1).The evidence that TIN-ag is a novel by centrifugation (2800 x g, 10 min), resuspended in 1.0 M NaCl in the member of the family of basement membrane glycoproteins is above cold inhibitor buffer, sonicated three times for periods of 2 min on ice, and washed by repeated centrifugation and resuspension (first with based on the following observations. First, antibodies against 1.0 M NaCl in inhibitor buffer, then with double distilled HZO). The otherbasementmembraneglycoproteinsdonotcross-react basement membrane prepared with this procedure wasincubated overwith TIN-ag. Second, antibodies raised against TIN-ag in ani- night at 37 “C with 1mg of collagenase (fromclostridium histolyticum, mals and circulating autoantibodies in patients do not crosscode: CLSPA, Worthington Biochemical Corp.) per 10 ml of wet tissue and brought to a volume of 40-100 ml with 50 m~ Tris-HC1,0.2 M NaC1, react with other basement membrane glycoproteins. Third, peptides derived from three different regions of the molecule 2.0 m~ CaC12, pH 7.4. Followingcollagenase digestion, the sample was exhibit amino acid sequences that do not show any similarity dialyzed against 2.0 M urea, 50 m~ Tris-HC1,0.2 M NaC1, 1.0 m~ CaC12, pH 7.4, and after centrifugation (10,000rpm for 10 min) and concenwith the amino acid sequences of known macromolecules (1). tration it was applied to a Sephacryl S-300 gelfiltration column (2.5 x Recently, a clone coding fora total of 173 amino acidshas been 85 cm). The fractions containing TIN-ag were pooled, concentrated with identified; comparisonof this sequence with known amino acid an Amicon concentrator (filter YM-lO), dialyzed against 2.0 M urea, 50 sequences has shown lack of similarities to any known protein m~ Tris-HC1,l.Om~ CaC12, pH 6.8,and then furtherpurified by cationexchange chromatography on a 25-ml S-Sepharose column equilibrated sequence.2 Immunofluorescence microscopy has revealed that TIN-ag in the same buffer. This column was developed by batchwise elution with 0.1 M NaCl followed by elution with 0.7 M NaCl in the column shows reactivity mainly in the kidney cortex and to a lesser buffer. The purified TIN-ag in the0.7 M NaCl elution of the column was aliquoted and stored at -80 “C. Laminin and type IV collagen wereextracted from EHS tumor grown * This work was supported by National Institutes of Health Grants DK 43569 (to A. S. C.) and DK 36007 (to R. J. B.) and by an American subcutaneously in mice made lathyritic by adding 0.1% p-aminoproHeart Association grant (to A. S. C.). The costs of publication of this pionitrile to the drinking water, according to protocols previously dearticle were defrayed in part by the payment of page charges. This scribed (3, 4). Laminin was stored in liquid nitrogen, and type IV colarticle must therefore be hereby marked “advertisement”in accordance lagen was stored on ice. Beforeuse, both macromolecules werecleared with 18 U.S.C. Section 1734 solely to indicate this fact. from large aggregates by Centrifugationfor 10 min at 10,000 rpm, $ To whom all correspondence should be addressed: Box 609 UMHC, unless stated otherwise. Dept. Lab. Med. Pathology, University of Minnesota Medical School, 420 Delaware St. S.E., Minneapolis, MN 55455. lZ5I Labeling of TIN-ag The abbreviations used are: TIN-ag, tubulointerstitial nephritis anTIN-ag was labeled with N a T (DuPont-New England Nuclear) by tigen; BM, basement membrane; rpm, revolutiondminute; EHS, Engelbreth-Holm-Swarm; PBS, phosphate-buffered saline; BSA, bovine se- chemical oxidationusing immobilized chloramine T (Iodo-Beads,Pierce rum albumin; cpm, countdminute; FBS, fetal bovine serum; DMEM, Chemical Co.) according to the manufacturer’ s instructions. Briefly, m CaC12, pH6.8 (reaction TIN-ag in 50 m~ Tris, 2 M urea, 0.5 M NaCl, 1x Dulbecco’s modified Eagle’s medium. buffer), was concentrated in aquacide.After the reaction of 5 Iodo-Beads R. J. Butkowski, unpublished observations.
1654
TIN Antigen Interactions with 1mCi of Na12'I in 150 pl of the reaction buffer for 5 min, approximately 100 pg of the protein in about 500 pl of reaction buffer wasadded to the reactivial. The reaction was allowed to proceed for 10 min. Finally, gel filtration through a 1 x 25cm Sephadex G-25-150 (Sigma) column was performed to remove excess Na12'I and unincorporated lZ6Izfrom the iodinated protein. Incorporated lZ5Iin TIN-ag was estimated by trichloroacetic acid precipitation. Solid-phase BindingAssays Binding assays were performed to quantify the binding of TIN-ag to laminin, type IV collagen, and heparin, following previously described methodolo& (5). Laminin and DpeN Collagen-TIN-agBinding Assays-Laminin or type IV collagen ai-60 p g / d in-PBS was coated0 2 0 96-well polystyrene Immulon 1plates (Dynatech Laboratories, Inc., Alexandria, VA), using 50 *well, and dried overnight at 29 "C. Under these conditions 1.8 ng of laminin3 and 1.3 ng of type IV (6) remained bound to each plastic well, as had previously been determined using radiolabeled proteins. Uncoated sites in theplastic wells were blockedwith 200 pVwell of 10 mg/ml BSAin PBS, pH 7.4 (blocking buffer), for 2 hat 37 "C.Unlabeled TIN-ag at various dilutions ranging from 300 to 0 pg/ml (final concentration) in blocking buffer was mixed with a standard amount of"'ITIN-ag (40,00~100,000 cpdwell, depending on the experiment), and 100 pl of this mixture was then transferred to washed laminin or type IV-coated wells and allowed to incubate for 3 h at 37 "C. The concentration of radiolabeled TIN-ag in this mixture was 1.5 pg/ml, and the wells of more than 150 p g / d cold TIN-ag were used to determine nonspecificbinding. This amount (-700 cpm, whichwas about 1% ofthe total counts added) was subtracted from the total binding before final calculations. Nonspecific binding was also estimated by measuring the binding of lZ5I-TIN-agto coated BSA and was found to be at a similar level. After incubation, the wells were washed thoroughly with 0.05% Triton-X-100in blocking buffer,and the bound protein was solubilized by incubation with 100 plof lysis buffer (0.5 M NaOH, 1%SDS in dH2O) for 30 min in a dry 60 "C oven. The amount of lZ5I in the lysate was then quantified in a gamma counter. TIN-ag-HeparinBinding Assays-Unlabeled TIN-ag, in triplicates of eight serial dilutions (from 192 pg/ml to 1.5 pg/ml) was mixed with 1251-TIN-agin PBS, and the mixture was coated in 96-well plates, 50 pVwell, and dried overnight at 29 "C. Under these conditions, in the range of concentrations used, 32% of the TIN-ag which was dried down was retained on the plastic well, as determined by the percentage of the bound radiolabeled TIN-ag. For the TIN-ag-heparin binding assays, TIN-ag at concentrations of 50 and 100 pg/ml in PBS was coatedin 96-well plates, using 50 pVwell, and dried overnight at 29 "C. As a control, plates were coatedwith BSA and EHS-laminin so that the final amount of protein adsorbed was equimolar with TIN-ag. The dried wells were then blocked with 200 buffer), for 2 hat 37 "C. pVwell of 2 mg/ml BSAin PBS, pH 7.4 (blocking After removal of this buffer, 100 pl of various concentrations of unlabeled heparin (from porcineintestinal mucosa, grade 1, Sigma), mixed with a standardamount of [3Hlheparin(from the same source, DupontNEN) in blocking buffer, wasadded to each well and incubated for 3 h a t 37 "C. Wells in which the ratio of unlabeled to labeled heparin was greater than 1OO:l were used to determine nonspecific binding. Unbound heparin was removed by washing three times with 200 @well wash buffer (0.05%Triton-X-100in blocking buffer), and bound heparin was lysed by adding 100 pVwe11 lysis buffer for 30 min a t 60 "C. 3H in the lysate was quantified in a Beckman LS-3801 scintillation counter. To exclude the possibility that denaturation of TIN-ag during drying blocks the ability to bind to heparin, an additional experiment was performed in which TIN-ag was coated in solution using 100 pVwell, overnight, a t 4 "C on a horizontal shaker. Solid-phase binding assays were performed in triplicates, at least four times each. n r b i d i t y Measurements In order to study the role of TIN-ag in the polymerization processof laminin and type IV collagen, the development of turbidity in solutions of these macromolecules in the presence or absence of TIN-ag was monitored. Samples were incubated in quartzcuvettes at 37 "C,and the temperature was maintained automatically using a Peltier 111kinetics System (Beckman). The change of absorbance a t 360 nm was followed with a Beckman DU-6 spectrophotometer. A. S. Charonis, unpublished data.
1655
Stored laminin was thawed and successively dialyzedat 4 "C against (a) 100 rm Tris-HC1, pH 7.4; ( b )0.5 M CaC12, 100 n m Tris-HC1, pH 7.4; ( c ) 100 rm Tris-HC1, pH 7.4; and ( d ) PBS. All solutions contained 50
pg/ml phenylmethylsulfonylfluoride. TIN-agwas also dialyzedagainst PBS. Laminin solution was centrifuged at 38,000 rpm for 20 min in a Beckman L8-M ultracentrifuge to clear large aggregates, and TIN-ag was centrifuged at 12,000 x g for 15 min in a Beckman microfuge 12. Aliquots of a constant amount of laminin (350 pg/nd) were mixed with increasing concentrations of TIN-ag in a final volume of 800 pl. In experiments where the reversibility of laminin polymerization was examined, several aliquots of laminin (each at 350 pglml in a volume of 600 pl) were allowed to develop turbidity for 40 min, after which the same volume (200 $) of TIN-ag solution a t various concentrations was added, and the change in the plateau value of turbidity was monitored. Q p e IV collagen was dialyzed against PBS containing 50 pg/ml phenylmethylsulfonyl fluorideand centrifuged at 15,000rpm for 30min at 4 "C. Aliquots of a constant amount of type IV collagen (350 pg/ml) were mixed withincreasing concentrations of TIN-ag in a final volume of 800 pl. Turbidity experiments were performed at least four times using different batches of isolated macromolecules.
Cell Adhesion Assays Two different cell types were used to test adhesion to TIN-ag: mouse tubular epithelial cells and bovine aortic endothelial cells. The mouse tubular epithelial cells were kindly provided by Dr. E. G. Neilson ( R e nal-Electrolyte Section, University of Pennsylvania, Philadelphia) as a long term cultured cell line (MCT epithelial cells transformed with a non-replicating, noncapsid-forming SV-40 virus (7). The cellswere maintained in RPMI-1640 (Sigma) containing 10%FBS,100pg/ml streptomycin, 100 IU/ml penicillin. The bovine aortic endothelial cells were isolated as described by Schwartz with a few modifications (8) and were cultured in DMEM, 10% FBS, 10% calf serum, 100 pg/ml streptomycin, 100 IU/ml penicillin. Experiments were carried out in 96-wellpolystyreneImmulon 1 plates. Laminin, type IV collagen, TIN-ag orBSA serially diluted in PBS was coated ontothe wells using 50 pVwell and dried overnight at 29 "C. The amount of each BM protein that remains adsorbed to the plastic wells under these conditions wasdetermined using radiolabeled proteins. In the range of concentrations used, protein adsorption onto the wells constitutes 60% of the input for laminin, 45% for type IV collagen, and 32% forTIN-ag. Cells in T-75 flasks with 80-90% confluence were incubated overnight in methionine-free DMEM (Life Technologies, Inc.) containing 10% FBS with 800 pCi of [35Slmethionine/flask.After discarding the radioactive medium and rinsing twice with Versene (PBS + 0.5 rm EDTA) (Life Technologies,Inc.), the cells wereharvested by adding 2 ml of 0.05% trypsin + 0.5 n m EDTAfor 2 min. The trypsinization was terminated by the addition ofDMEM + 10% FBS, and the cells were washedtwice with serum-free DMEM and resuspended to 50,000 celldnd in binding buffer (DMEM, 25 II~MHEPES, 2 mg/ml BSA, pH 7.4). 100 pl of the suspension was plated in the precoated wells which had just been blockedwith 200 pVwell of blockingbuffer (PBS, 2 mg/ml BSA, pH 7.4) for2 h at 37 "C. The plates were incubated for 90 min at 37 "C. Next, nonadherent cells were removed by aspiration, and the wells were washed three times with 200 p1 of blocking buffer/well. Adherent cells were solubilized with 100 p1of 0.5 M NaOH + 1%SDS and quantified in a scintillation counter. Percent adhesion was calculated as thepercent of counts added to each well that remained bound to the well at the end of the assay. Cell adhesion assays were performedin triplicate wells at least three times for each cell line. RESULTS
In order to examine and quantify the interaction between TIN-ag and other basement membrane macromolecules, solidphase binding assays were performed by coating laminin, type lV collagen, and TIN-ag on 96-well plastic plates. For the interaction between TIN-ag and laminin, EHS-derived laminin wascoated as described; under theseconditions, it was determined that 2.0 pmol of laminin was retained on each well (data not shown). First, the binding of increasing concentrations of radiolabeled TIN-ag was examined and found to be saturable (Fig. L4). Then, the specificity of this interac-
TIN Antigen
1656
Interactions
&
0
50
100 TINsag
B a m
1000
9
600
150
200
added
250
300
? i
0.6
< B
0.4
B
0.2
50
350
300
350
000
-200 10
0
centrations (M-320
TIN-ag]
interaction.
coated
tug/ml)
A,
on 96-well
saturable plates.
binding
Increasing
of con-
pmol/lOO pl) of a mixture of labeled and unlabeled
TIN-ag were added to quadruplicate wells coated with 2.2 pmol laminin and allowed to react for 3 h at 37 “C. The bound TIN-ag displayed per picomole of laminin coated. Inset, Scatchard analysis
of is of
solid-phase binding assay between TIN-ag and laminin; linear regression fits the data in a line with r = 0.96 (p < 0.001). *Bound TIN-ag has been expressed as moles per mole of laminin coated. Data are pooled from two different experiments. B, competition of binding of ‘x51-TIN-ag in solution to laminin coated on plastic wells by unlabeled TIN-ag. A
constant amount of 1261-TIN-ag (35,000 cpm, corresponding to 1.0 pg/ ml) was mixed with increasing the mixture allowed to react
concentrations with the coated
of unlabeled TIN-ag, and laminin for 3 h at 37 “C.
Unbound TIN-ag was removed by washing. Nonspecific binding of rz51TIN-ag to laminin in the presence of a loo-fold excess of unlabeled TIN-ag has been subtracted and the counts/mm of bound lz51-TIN-ag against
50
100 [cold
[cold
FIG. 1. Laminin-TIN-ag TIN-ag in solution to laminin
was plotted
100 150 200 250 TIN-ag addad (pm/wall)
(pm/well)
concentrations
of the competitor
“cold”
TIN-ag.
tion was tested by co-incubating a constant amount of radiolabeled TIN-ag with increasing concentrations of unlabeled TINag. The binding could be competed, as shown in Fig. IS. Scatchard analysis indicated a single line of linear regression with correlation of F = 0.96. The plot also suggested that approximately two binding sites for TIN-ag may exist on the laminin molecule and that they may exhibit a similar affinity with a Kd of 1.2 pi (Fig. IA, inset). For the interaction between TIN-ag and type IV collagen, EHS-derived collagen was coated under the conditions described above. Under these conditions, it was determined that 2.6 pmol of type IV collagen was retained on each well (data not shown). The interaction between coated type IV collagen and increasing concentrations of radiolabeled TIN-ag was studied next. This interaction was found to be saturable, and at saturation TIN-ag interacted with type IV in a 1:l molar ratio (Fig. 2A). The binding was shown to exhibit specificity, since increasing concentrations of unlabeled TIN-ag could effectively compete the interaction between radiolabeled TIN-ag and type IV collagen (Fig. 2B). Scatchard analysis showed a single line of linear regression with a correlation of F = 0.95 and approximately one binding site/collagen molecule with a Kd of 1.0 pM (Fig. 2A, inset). In order to explore possible interactions between TIN-ag and the side chain of basement membrane proteoglycans, heparin was used as a model molecule. Solid-phase competition assays
150 TIN-sg]
200
250
300
(w/ml)
FIG. 2. Type IV collagen-TIN-ag interaction. A, saturable binding of TIN-ag in solution to type IV collagen coated on 96-well plates. Increasing concentrations (6.0-340 pmol/lOO pl) of TIN-ag were added to quadruplicate wells coated with 2.4 pmol of type IV collagen and allowed to react for 3 h at 37 “C. Bound TIN-ag is displayed per picomole of type IV collagen coated. Inset, Scatchard analysis of the solidphase binding assay between TIN-ag and type IV collagen; linear regression fits the data in a line with r = 0.95 (p < 0.001). *Bound TIN-ag has been expressed as moles per mole of type IV collagen coated. Data are pooled from three different experiments. B, competition of binding of ‘x61-TIN-ag in solution to type IV collagen coated on plastic wells by unlabeled TIN-ag. A constant amount of rz51-TIN-ag (55,000 cpm, corresponding to 2.0 pg/mll was mixed with increasing concentrations of unlabeled TIN-ag, and the mixture allowed to react with the coated type IV collagen for 3 h at 37 “C. Unbound TIN-ag was removed by washing. Nonspecific binding of iz51-TIN-ag to type IV collagen in the presence of a loo-fold excess of unlabeled TIN-ag was subtracted and counts/
minute of bound rz61-TIN-ag was plotted against concentration of the competitor
“cold”
TIN-ag.
were performed in which either 14 or 28 pmol of TIN-ag or equimolar amount of laminin was retained in each well. Laminin was used in these experiments as a positive control. The binding of tritiated heparin was measured, and the ability of increasing concentrations of unlabeled heparin to compete for this binding was monitored (Fig. 3). Under the experimental conditions used, in which all solutions were at physiologic ionic strength, heparin interacted with laminin in a specific manner, but no specific interaction with TIN-ag was observed. To exclude the possibility that denaturation of TIN-ag during coating reduced the ability to interact with heparin, a different experimental approach was used in which TIN-ag was coated in solution at a concentration of 100 &ml and not allowed to dry. This approach also showed no specific interaction between heparin and TIN-ag (data not shown). Collectively, the above observations suggest that TIN-ag interacts in a specific manner with laminin and type IV collagen and that this interaction exhibits relatively low affinity for both macromolecules. The data also do not exclude the possibility that interactions may exist between TIN-ag and basement membrane-associated or cell surface-associated proteoglycans at sites other than the negatively charged side chains. The ability of TIN-ag to interact with laminin and type IV collagen raised the possibility that TIN-ag may, to some extent, interfere its role
with
the
process
in the self-assembly
of their
of laminin
polymerization.
Therefore,
and type IV collagen was
TIN Antigen Interactions '
2000-t
'
''.,..I
'
'
"l.l.I
'
'
0.05
" " "
U
a
15001-
8
1000
'0
8
>
0
2 -
"Z
1657
!
j 0. 500
1
,
a
E
O.O1
0 4
"
10
100
0
1000
10
20
FIG.3. Absence of heparin binding to TIN-ag dried onto plastic wells. Wells coated with 14 pmol of TIN-ag or laminin were incubated for 3 h at 37 "C with a standard amount ofL3H1heparin and various concentrations of unlabeled heparin. Heparin interacted with laminin (0) in a specific manner but did not interact with TIN-ag (0).
40
50
60
0.07 0.06
m
studied by following over time the change in turbidity of solutions containing thesemacromolecules. Laminin at a final concentration of 350 pg/ml in PBS was incubated at 37 "C for 60 min in theabsence or the presence of increasing TIN-ag concentrations, and thedevelopment of turbidity was monitored at AaeO.The presence of TIN-ag resulted in a decrease in the plateau values, although the kinetics of polymer formation weresimilar (Fig. 4A). The decrease of maximal turbidity correlated with the amountof TIN-ag present in the incubation mixture ina dose-dependent manner: at equimolar concentrations, TIN-agsuppressed themaximal turbidity of laminin by 30%.At higher molar ratios the decrease was even more dramatic: 66% decrease was observed when TIN-ag was present 8-fold in molar excess. TIN-ag aloneat the highest concentration used did not show any development of turbidity (Fig. 4A). Thisapproach examined the effect of TIN-ag on the process of laminin polymerization but did not provide information on the effect of TIN-ag on already preformed laminin aggregates. To study this, laminin was first allowed to polymerize as described, and when the plateauvalue was reached, various amounts of TIN-ag were added to the laminin solution and thechange in the turbidity was monitored over time. It was observed that TIN-ag was able to reduce the plateau value(monitored after theaddition of PBS without any TIN-ag) ina concentration-dependent manner (Fig. 4B ). At the highest concentration of TIN-ag used (molar ratio of TIN-ag/ laminin of 9:1),the drop in the plateau value was34%. Type IV collagen at a final concentration of 350 pg/ml in PBS was also incubated at 37 "C for 60 min in the absence or the presence of increasing concentrations of TIN-ag. Type IV collagen alone showed an increase in turbidity, as expected. However, in contrast with laminin, the presence of TIN-ag in the incubation mixture did not have any effect on the kinetics or the plateau valueof the turbidity of type IV collagen (Fig. 5). The above observations suggest that the TIN-ag-binding site(s) on laminin should be localized to domain(s) involved in the process of polymerization, whereas thebinding site on type IV collagen should be at a position that is not critical for its ability to polymerize. In order to begin understanding the influence that TIN-ag might exert on the cellular phenotype, experiments were designed to examine the possible role of TIN-ag in promoting cell adhesion. In these experiments,plastic plates were coated with TIN-ag. Coated laminin and type IV collagen were used as positive controls, whereas BSA served as a negative control. The exact amount of each macromolecule retained on the well was precisely calculated. The adhesion of tubular epithelial and aortic endothelial cells was examined as described above. The data are presented in Fig. 6A for tubular epithelial cells
30
tlme(min)
[cold heparin] (pg/ml)
0 0
0.05 0.04
0.03
a 0.02 0.01 0
0
10
20
30
40 SO 6 0 time (mln)
70
80
90
FIG.4. A, effect of TIN-ag on laminin turbidity. Laminin at 350 p g / d in PBS was incubated at 37 "C for 60 min in the absence or presence of increasing TIN-ag concentrations. The change in absorbance at 360 nm was plotted against time. 0, laminin alone; x, 20 pg/ml TIN-ag added; 0, 40 pg/ml TIN-ag added; 0, 80 pg/ml TIN-ag added; D, 160 pg/ml TIN-ag added; e, TIN-ag alone (160 pg/ml). B , effect of TIN-ag on preformed laminin aggregates. 600-p1 aliquots of laminin a t 350 pg/ml in PBS were incubated at 37 "C until a plateau value was reached (40 min). At this time, 200 pl of PBS with TIN-ag at various final concentrations (150 pg/ml (Dl, 50 pg/ml (x), or PBS alone (0) as a control)were added to the turbid solution, and theA3mwas followed for50 additional min. The solid circles (0)show the turbidity of laminin at 350 pg/ml for 90 min.
-
/
0.005
0
10
20
30
40
50
60
time (mln)
FIG.5 . Effect of TIN-ag on type IV collagen turbidity.Type N collagen at 350 pg/ml in PBS was incubated at 37 "C for 60 min in the absence or presence of increasing TIN-ag concentrations. The change in absorbance at 360 nm was plotted against time. D, type IV collagen alone; 0 , 4 0 pdml TIN-ag added; 0 , 2 0 0 pg/ml TIN-ag added; 0,TIN-ag alone (200 pg/ml).
and in Fig. 6B for aortic endothelial cells. It can be observed that for both cell types, TIN-ag is a strong adhesive macromolecule, although not as strong as laminin or type IV collagen. In the case of tubular epithelial cells, TIN-ag was only one order of magnitude lessadhesive compared on a molar basis to laminin and type IV collagen; however, in the case of aortic endothelial cells, TIN-ag was 2 orders of magnitude less adhesive than type IV collagen or laminin. Data similar to those presented in Fig. 6, A and B , were generated in experiments in which the four types of macromolecules were coated in the wells in solution and not allowed to dry (data notshown).
Interactions laminin polymers resulted in a dose-dependent decrease of the existing plateau values, indicating that the interaction between TIN-ag and laminin can exert a "dissolving" effect on existing laminin polymers. It is established that such laminin networks exist in situ (9); therefore, it is conceivable that the presence of high TIN-ag concentrations locallycouldaffect these networks and consequently the overall structure of the basement membrane at specific sites. So far, only heparin has been observed to be able to regulate laminin self-association (10,111; TIN-ag is thesecond macromoleculethat is reported to regulate laminin polymerization. In contrast to this effect on 0.01 0.1 1 10 1 oz laminin, TIN-ag didnot exert any effect on the ability of type IV pmoles of adsorbed protein/well collagen to polymerize. The experiments presented here were performedusing EHS80 extracted laminin and type IV collagen. These two major base70 ment membrane glycoproteins exist in various isoforms (12) 60 exhibiting differential tissue localization. Therefore, the possiE .P 5 0 bility cannot be excluded that TIN-ag could exert more (or less) 1 pronounced effectson the binding or the polymerization of spe6P 40 cific isoforms of laminin and type N collagen. L 30 In turbidity experiments, TIN-ag alone did not show any 20 to polymerize. This was confirmed at the electron tendency 10 microscopic levelwith rotary shadowing. Using this technique, 0 no aggregation of TIN-ag was ~ b s e r v e d . ~ 0.01 0.1 1 10 1 o2 Most basement membrane macromolecules studied so far pmoles of adsorbed proteln/well have been able to promote cellular adhesion (€413-15)The role FIG.6. A, adhesion of tubular epithelial cells to TIN-ag. Mouse tubu- of TIN-ag in promoting cell adhesion was studied using two lar epithelial cells metabolically labeled with [35Slmethionine, were plated (5 x lo3 celldwell) into wells coated with increasing concentra- different cell types. Evidence is presented that TIN-ag protions of type IV collagen (0);laminin (W); TIN-ag (0);or BSA (0). After motes adhesion of kidney tubular epithelial cells and aortic 90 min the nonadherent cells were washed away, and the percentage of endothelial cells. Interestingly TIN-ag was found to be more adhesion was determined by counting the remaining radioactivity. Each adhesive for tubular epithelial cells compared to aortic endopoint represents the mean of three measurements. B , adhesion of aortic endothelial cells to TIN-ag. Bovine aorticendothelial cells metabolically thelial cells. More work needs to be done to explore the possilabeled with [35Slmethioninewere plated (5 x lo3 celldwell) into wells bility that TIN-ag promotes cell adhesion in a tissue- or cellcoated with increasing concentrations of type IV collagen (0);laminin specific manner. An important aspect of this effort will be to (W); TIN-ag (0);or BSA (0). After 90 min the nonadherent cells were identify cell surface macromolecules that interact specifically washed away, and the percentage of adhesion was determinedby counting the remaining radioactivity. Each point represents the mean of with TIN-ag. Given the promiscuity of interactions between integrins and matrixmacromolecules,it is conceivable that one three measurements. or more of the known integrin heterodimers mediate cell surIn summary, cell adhesion assays establish that TIN-ag is an face interactions with TIN-ag. Lack of ionic interactions with adhesion-promoting macromolecule. However, further experi- proteoglycan side chains (see above) suggests that cell surface ments will be required to correlate strength of adhesion with proteoglycans might not mediate cellular interactions with different cell types and to analyze the macromolecular compo- TIN-ag. However, weak multiple interactions with side chains or interactions with the protein core of cell surface proteoglynents on the cell surface that mediate this response. cans cannot be excluded. DISCUSSION Immunohistochemical localizationprovidedevidence that In this report, the interactions between TIN-ag, a novel base- among all the tissues examined, reactivity for TIN-ag is highest ment membrane macromolecule, and other basement mem- in the basement membrane underlying kidney tubular epithebrane components werestudied. In addition, the role of TIN-ag lial cells and intestinal epithelial cells of the ileum (2).Both these cell types are involved in extensive absorption from their in promoting cell adhesion was explored. TIN-ag was found to interact specifically with laminin and apical surface, which faces the lumen, and they release most type IV collagen. The Kd of this interaction is in the low micro- absorbed material to their basolateral surface from which acmolar range. It should be kept in mind that this affinity is cess to the microcirculation (capillaries, lymphatics) is gained calculated using solid-phase binding assays where macromol- after crossing the epithelial basement membrane. It is intriguecules are dried on plastic wells. Consequently, possibledena- ing to speculate that the presence of TIN-ag in large amounts turation and steric alterations of putative binding sites may at these two sites might be related to functional aspects of the result inunderestimation of the strength of these interactions. epithelial cells and that its presence regionally modifies the Heparin was used as a model system in order to study the structure of the basement membrane, thus creating amicrodointeractions between TIN-ag and highly negatively charged main supporting functional requirements. proteoglycan side chains. Under the experimental conditions Acknowledgment-We thank Dr. Eric Neilson (University of Pennused (solid-phase assays after dry or wet coating using buffers sylvania, Renal-Electrolyte Section) for providing the MCT epithelial of physiologic ionic strength), no specific interaction was de- cells. tected. Although such ionic interactions cannot be excluded, it REFERENCES is likely that they are very weak, if present at all. The ability of laminin to self-associate was affected by the 1. Butkowski, R. J., Langeveld,J. P., Wieslander, J.,Brentjens,J. R., andAndres, G . A. (1990)J . Biol. Chem. 285,21091-21098 presence of TIN-ag. Increasing concentrations of TIN-ag re2. Butkowski, R. J., Kleppel,M. M., Katz, A,, Michael,A. F., and Fish,A. J. (1991) sulted in decreasing plateau values, indicating impairment of Kidney Int. 40,83-46 3. Timpl, R., Rhode, H., Gehron-Robey, P., Rennard, S. I., Foidart, J.-M., and polymerization. Furthermore, addition of TIN-ag to preformed
TIN Antigen Interactions Martin, G. R. (1979)J. Bwl. Chem. 264,9933-9937 4. Kleinman, H. K., McGarvey, M. L., Liotta, L. A,, Robey, l?
G., ‘Ayggvason, K.,
Martin, and G . R. (1982) Biochemistry 21,6188-6193 5. Skubitz, A. P. N., McCarthy, J. B., Charonis, A. S., and Furht, L.T.(1988) J. Bioi. Chem. 263,48614868 6. Tsilibary, E.C., Koliakos, G . G .Charonis, , A. S., Vogel, A. M., Reger, L. A., and Furcht, L. T. (1988) J . Bioi. Chem. 1988, 19112-19118 7. Haverty, T. P., Kelly, C. J., Hines, W. H., Amenta, P. S., Watanabe, M., Harper, R.A., Kefalides, N.A., and Neilson, E. G. (1988)J. Cell B i d . 107,1359-1368 8. Herbst, T. J., McCarthy, J. B., Tsilibary, E. C . , and Furcht, L. T.(1988)J. Cell Bioi. 106. 1365-1373 9. Yurchenco, P. D.,Cheng, Y. S . , and Colognato, H. (1992) J. Cell Biol. 117,
1659
1119-1133 10. Kouzi, K. K , Koliakos, G . G., Tsilibary, E. C., Furcht, L. T.,and Charonis, A. S. (1989) J. Bwi. Chem. 264,17971-17978 11. Yurchenco, P. D., Cheng, Y. S., and Sehittny, J. C. (1990) J. Biol. Chem. 286, 3981-3991 12. Sanes, J. R., Engvall, E., Butkowski, R., and Hunter, D. D. (1990)J. Cell Bioi. 111, 1685-1699 13. Paulsson, M.(1992) Crit. Rev. Biochem. Mol. Bioi. 27,93-127 14. Carlsson, R., Engvall, E., Freeman, A,, and Ruoslahti, E. (1981) Proc. Nati. Acad. Sei. U. S. A. 78, 2403-2406 15. Aumailley, M., Nurcombe, V., Edgar, D., Paulsson, M., and Timpl, R. (1987)J. Biol. Chem. 282,11532-11538
