Proc. Natl. Acad. Sci. USAVol. 93, pp. 2915-2919, April 1996Biochemistry

Recombinant human uteroglobin suppresses cellular invasivenessvia a novel class of high-affinity cell surface binding site

(Clara cell 10-kDa protein/invasion/extracellular matrix/metastasis)

GOPAL C. KUNDU, GIUDITrA MANTILE*, LUCIO MIELE, ELEONORA CORDELLA-MIELE, AND ANIL B. MUKHERJEEtSection on Developmental Genetics, Human Genetics Branch, National Institute of Child Health and Human Development, National Institutes of Health,Bethesda, MD 20892-1830

Communicated by V. Ramalingaswami, All India Institute of Medical Sciences, New Delhi, India, December 5, 1995

ABSTRACT The mechanism(s) that regulates invasion oftrophoblasts through the uterine epithelium during embryoimplantation and nidation in hemochorial placental mam-mals is poorly understood. While limited trophoblast invasionis essential for the establishment of normal pregnancy, dys-regulation of this process may contribute to the pathogenesisof choriocarcinoma, a highly invasive and lethal form ofcancer arising from the trophoblasts. We have previouslydemonstrated that rabbit uteroglobin (UG), a cytokine-like,antiinflammatory protein, produced by the endometrial epi-thelium during pregnancy, has a potent antichemotactic effecton neutrophils and monocytes in vitro. Here, we report thatrecombinant human UG (hUG) dramatically suppresses in-vasion of human trophoblasts and NIH 3T3 cells through anartificial basement membrane (Matrigel) in vitro but has noeffect on that of human choriocarcinoma cells. We identifieda previously unreported high-affinity, high molecular weight('190 kDa), nonglycosylated hUG-binding protein, readilydetectable on human trophoblasts and NIH 3T3 cells buttotally lacking on choriocarcinoma cells. Taken together,these results raise the possibility that (i) hUG plays a criticalrole in regulating cellular invasiveness, at least in part, via itspreviously unrecognized cell surface binding site, and (ii)some of the numerous biological activities of proteins of the UGfamily, reported so far, may be mediated via this binding site.

Cellular migration and invasion of the extracellular matrix areessential for embryonic development, inflammation, repair ofinjured tissues, and metastasis of distant organs by malignantcancer cells. During pregnancy in hemochorial placental mam-mals (e.g., mice and humans), the trophoblasts spearheadembryonic invasion through the uterine epithelium deep intothe stroma. During this process, these cells penetrate theextracellular matrix (i.e., the basement membrane) for implan-tation and nidation to take place (1). This invasion of tropho-blasts is a prerequisite for the establishment of successfulpregnancy (reviewed in ref. 2). Although trophoblast invasionof the endometrium in normal pregnancy is limited andcritically regulated, these cells are moderately invasive inhydatidiform molar pregnancy and are highly invasive inchoriocarcinoma, a rare but one of the most lethal forms ofcancers arising in the reproductive organs. Despite muchprogress in the general field of trophoblast biology, themechanism(s) that regulates the invasiveness of these cells isstill not clearly understood.

Rabbit blastokinin (3) or uteroglobin (UG; ref. 4) is asteroid-dependent, multifunctional, cytokine-like protein withpotent antiinflammatory activity (for review, see refs. 5-7).This protein was first discovered in the rabbit uterus duringearly pregnancy and subsequently found in many other non-reproductive organs. Interestingly, the highest level of UG

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

expression occurs in the rabbit (3, 4) and human endometrium(8) during the progesterone-dominated phase of the ovarianmenstrual cycle, when implantation of the embryo in theuterus takes place. Prolactin appears to further stimulateprogesterone-induced UG gene expression in the rabbit uterus(9). Human UG (hUG; refs. 10-12), or Clara cell 10-kDaprotein (13-15), is the counterpart of rabbit UG (16-19). BothhUG and UG effectively inhibit (15, 16, 20-23) the activitiesof low molecular weight group I and group II extracellularphospholipases A2 (E.C. 3.1.1.4), a class of key enzymes thatplay critical role(s) in the production of proinflammatory lipidmediators (reviewed in refs. 24-26). We previously demon-strated that UG is a potent inhibitor of neutrophil andmonocyte chemotaxis (6, 27). Robinson et al. (28) havereported an active transport of UG into the blastocoele of thepreimplantation rabbit embryos, through the trophoblast layervia a carrier-mediated process. As mentioned above, tropho-blast cell migration and invasion through the endometrialepithelium is an essential prerequisite for the establishment ofpregnancy. However, this invasion is a precisely regulatedprocess, as the trophoblasts do not invade extrauterine organs.Since the peak level of UG and hUG gene expression in theuterus coincides with the time of implantation and a relativelylower level of expression in this organ is maintained through-out gestation Peri et al. (29), we sought to determine: (i)whether hUG had any effect on the invasiveness of humantrophoblast cells in vitro, and (ii) if hUG exerted this effect viaa specific interaction with human trophoblast cells and, if so,(iii) whether this cell surface interaction and concomitantbiological effects of this protein are species specific.

Here, we report that: (i) hUG dramatically suppressesinvasion of an artificial basement membrane (Matrigel) byhuman trophoblasts and NIH 3T3 cells; (ii) hUG fails to exertthis effect on human choriocarcinoma cells; (iii) recombinanthUG binds to human trophoblasts and NIH 3T3 cells, sug-gesting a lack of species specificity; (iv) the results of affinity-crosslinking of 125I-labeled hUG (125I-hUG) as the ligandindicate that a single-class, high-affinity, high molecular weight(-190 kDa), nonglycosylated binding protein is present onhuman trophoblasts and NIH 3T3 cells; and (v) although thisbinding is readily detectable on human trophoblast and NIH3T3 cell surfaces, it is totally absent from those of the humanchoriocarcinoma cells. These results raise the possibility thathUG plays an important and critical physiological role inregulating cellular invasiveness, at least in part, via a high-affinity, cell surface binding site.

MATERIALS AND METHODSRadioiodination of Recombinant hUG. Recombinant hUG

was expressed in Escherichia coli and purified to homogeneity

Abbreviations: UG, uteroglobin; DSS, disuccinimidyl suberate.*Present address: Gerontology Research Center, National Institute onAging, National Institutes of Health, Baltimore, MD 21224.tTo whom reprint requests should be addressed.

2915

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996)

by the method of Mantile et al. (16). This protein (20 p,g) wasradioiodinated by using Na125I (2 mCi; carrier free; 1 Ci = 37GBq) and Iodo-Beads. The reaction was carried out in 130 tLlof phosphate-buffered saline (PBS; pH 7.4) at 25°C for 10 min,and 125I-hUG was purified by Sephadex G-25 spun-columnchromatography (1200 x g for 4 min). The column was

prewashed with 0.05% bovine serum albumin (BSA) in PBS.The specific activity of purified carrier-free 125I-hUG was 21tuCi/ptg.

Cell Cultures. Human trophoblasts [transformed with a

replication-defective, temperature-sensitive (ts) mutant of anadenovirus-simian virus 40 chimeric virus (Adeno-SV40)](30) and two human choriocarcinoma cell lines (JAR andJEG) were grown in a-modified Eagle's medium supple-mented with 4% (vol/vol) fetal calf serum (GIBCO), 100 unitsof penicillin per ml, 100 ,jg of streptomycin per ml, and 10 mMglutamine in a humidified atmosphere of 5% C02/95% air.Human trophoblasts were cultured either at 33°C or 39°C, andhuman choriocarcinoma cell liens were grown and maintainedat 37°C. NIH 3T3 cells were grown in Dulbecco's modifiedEagle's medium (DMEM) supplemented with 10% fetal calfserum (GIBCO), 100 units of penicillin per ml, 100 pAg ofstreptomycin per ml, and 10 mM glutamine in a humidifiedatmosphere of 5% C02/95% air at 37°C.Chemoinvasion Studies. Chemoinvasion assays were per-

formed as described (31). Briefly, confluent cells were har-vested with 0.05% trypsin and 0.02% EDTA and then centri-fuged at 800 X g for 10 min. Cells were washed with trypsin-neutralizing solution (Clonetics) followed by DMEMcontaining 0.1% BSA and were resuspended in DMEM con-

taining 0.1% BSA. The lower compartment of the invasionchamber was filled with fibroblast-conditioned medium(FCM), which was used as a chemoattractant. FCM was

prepared by incubating proliferating cultures of NIH 3T3fibroblasts for 24 hr. The lower compartment was overlaid with8-ptm pore size polyethylene phthalate membrane precoatedwith Matrigel basement membrane matrix at 125 ptg/cm2. Thecells (1.6 x 105 per well) were seeded in the upper compart-ment of the prehydrated Matrigel-coated invasion chambers inthe absence or presence of 1 tLM hUG and were incubated at33°C for 36 hr (human trophoblasts) or at 37°C for 36 hr(human choriocarcinoma and NIH 3T3 cells) in a incubatorwith humidified 5% C02/95% air. The cells that invaded theMatrigel and attached to the lower surface of the filter werestained with Giemsa for 3 min. The upper surface of the filterwas scraped with moist cotton swabs to remove Matrigel andnonmigrated cells. The chamber was washed three times withwater, the migrated cells were counted under an invertedmicroscope, and photomicrographs (x120) were taken byusing a Zeiss photomicroscope, Axiovert 405 M. The percent-age of cell invasion was determined by assuming the numberof cells that invaded in the absence of hUG as 100%.

Binding Studies. Cells, grown to confluence, in 12-wellplates, were washed once with cold PBS (pH 7.4) and thenincubated with 1.5 nM 125I-hUG in 1 ml of Hanks' balanced saltsolution (HBSS) (pH 7.6) containing 0.1% BSA in the absenceor presence of various concentrations of unlabeled hUG (1 pMto 1 ptM) at 4°C for 2 hr. The reaction was stopped by rapidremoval of unbound 125I-hUG, and the cells were washedthrice with cold PBS (pH 7.4) and solubilized in 1 M NaOHfollowed by addition of an equal volume of 1 M HC1. Theradioactivity was measured by using a y counter (ICN, model10/600 plus) with a counting efficiency of =80%.

Affinity Crosslinking Experiments. Confluent cell culturesin six-well plates, were washed with ice-cold PBS (pH 7.4) andincubated with 3.0 nM 125I-hUG in 2.0 ml of HBSS (pH 7.6)containing 0.1% BSA in the absence or presence of unlabeled1 ptM hUG for 2 hr at 4°C. After two washings with ice-coldPBS, the cells were incubated further with 0.20 mM disuccin-imidyl suberate (DSS) in 2.0 ml of HBSS (pH 7.6) for 20 min

at 4°C. The reaction was terminated by adding 50mM Tris-HClbuffer (pH 7.5), and cells were scraped, collected by centrif-ugation at 10,000 x g for 15 min, and lysed in 60 Aul of 1%Triton X-100 containing 1 mM phenylmethylsulfonyl fluoride,20 ,ag of leupeptin per ml, and 2 mM EDTA (lysing buffer).The supernatants (30 ptl) obtained by centrifugation at 10,000x g for 15 min were suspended in sample buffer in the presenceof 5% 2-mercaptoethanol, boiled for 5 min, and electropho-resed on a 4-20% gradient sodium dodecyl sulfate (SDS)/polyacrylamide gel (Bio-Rad). The gels were briefly stainedwith Coomassie blue, dried in a Bio-Rad gel dryer, andautoradiographed with Kodak X-Omat AR x-ray film.

Deglycosylation Studies. For pretreatment with glycosi-dases, confluent cells grown in six-well plates were scraped andcollected by centrifugation at 10,000 x g for 15 min, washedwith 100 mM Tris-HCl (pH 7.0), and incubated with eitherN-glycosidase F (3 units/100 Al) or BSA-free O-glycosidase (2milliunits/100 ual) in 100 t1l of 100 mM Tris-HCl buffer (pH7.0) at 37°C for 16 hr (32). The cells were collected bycentrifugation, washed with HBSS (pH 7.4), and incubatedfurther with 3.0 nM 125I-hUG in 500 pl of HBSS (pH 7.6)containing 0.1% BSA in the absence or presence of unlabeledhUG at 4°C for 2 hr. The cells were treated with 0.20 mM DSSfor crosslinking and lysed in 60 pul of lysing buffer. Thesupernatants collected after centrifugation, were mixed withsample buffer containing 5% 2-mercaptoethanol, boiled for 5min, electrophoresed on 4-20% gradient gel, and autoradio-graphed. For posttreatment with glycosidases, confluent cellswere incubated with 3.0 nM '25I-hUG for binding andcrosslinked with 0.20 mM DSS. The cells were scraped,washed, and boiled for 5 min in 30 ptl of 100 mM Tris-HClbuffer (pH 7.0) containing 0.5% SDS for denaturation. Thesample was diluted to 150 ptl with 100mM Tris-HCl buffer (pH7.0; final SDS concentration 0.1%) containing 0.5% N-octylglucoside, 0.4 mM phenylmethylsulfonyl fluoride, 10tLg/ml leupeptin, 2 mM EDTA and 1% 2-mercaptoethanol andincubated with N-glycosidase F (3 units/100 p.l) or BSA-freeO-glycosidase (2 milliunits/100 ual) at 37°C for 16 hr. Thesamples were centrifuged at 10,000 x g for 15 min, and theresulting supernatants were lyophilized, giving solids that wereresuspended in sample buffer; this mixture was boiled, elec-trophoresed, and autoradiographed.

RESULTS AND DISCUSSIONThe highest level of UG gene expression occurs in the humanendometrium during the midluteal phase of the ovarian men-strual cycle (8, 15), a critical period when implantation of theembryo in the uterus takes place. Because UG has beenreported to inhibit formyl-Met-Leu-Phe-induced chemotaxisof neutrophils and monocytes (6, 27), possibly via a receptor-mediated pathway, we determined whether hUG had anyeffect on the ability of human trophoblasts, human choriocar-cinoma, and NIH 3T3 cells to invade through a layer ofartificial basement membrane (Matrigel). The results of theseexperiments demonstrated that hUG causes a dramatic, dose-dependent (data not shown) suppression of invasion of Ma-trigel by these cells. Representative results of typical invasionexperiments with human trophoblasts (18% invasion) and NIH3T3 cells (21% invasion) are shown in Table 1 and Fig. 1 Top(a/b) and Bottom (elf), respectively. Most significantly, hUGhad no antiinvasive effect Fig. 1 Middle (c/d) on humanchoriocarcinoma cells (100% invasion). A nonspecific protein,myoglobin, had no suppressive effect on these cells, and whenunconditioned medium was used as chemoattractant in thelower chamber no invasion occurred.To determine the mechanism(s) of suppression of cellular

invasiveness, we first sought to delineate whether hUG had anyspecific binding on the cell surface. Thus, we made an initialsurvey of several cell types, which indicated that hUG-binding

2916 Biochemistry: Kundu et al.D

ownl

oade

d by

gue

st o

n A

ugus

t 20,

202

1

Proc. Natl. Acad. Sci. USA 93 (1996) 2917

.44hC3

. 4p~t''oa1,~~~~~~

;.z· ';

4rB

*

. e , * * ,

b%, ,'.. -* , ';§ S " > ;( % F¢f`* sk . "· X? .*s 4 < °

FIG. 1. Photomicrographic montage illustrating the effect of hUGon chemoinvasion of Adeno-SV40 virus-transformed normal humantrophoblast, human choriocarcinoma and NIH 3T3 cells. (Top) Humantrophoblasts: control (a) or with 1 ,uM hUG (b). (Middle) Humanchoriocarcinoma cells (JAR): control (c) or with 1 ,uM hUG (d).(Bottom) NIH 3T3 cells: control (e) or with 1 ,uM hUG (f).

is detectable only on very selected cell types. In fact, from a

panel of seven different cell types tested so far, we found onlyhuman trophoblasts and NIH 3T3 cells to be positive for125I-hUG binding. Therefore, we first characterized hUGbinding on human trophoblast cells. Since homogeneous pop-ulations of trophoblast cells from intact term placenta are

difficult to obtain, we used an established human trophoblastcell line. The growth and differentiation of these trophoblastsused in our investigation are temperature-dependent, as thiscell line was established by transformation with a temperature-sensitive, replication-defective mutant of a chimeric, Adeno-SV40 virus (30). These cells proliferate at a rapid rate at thepermissive temperature (33°C) and differentiate at the non-

permissive temperature (39°C). Thus, we carried out thebinding studies on these cells which were cultured either at33°C or at 39°C. Scatchard analysis of steady-state binding of'25I-hUG suggests the presence of a single class of specific cell

A

Cc-

O

a)

0o

Q)

E0

e--

om

1.2r

0.8l

0.41

0 8 16

surface binding sites with a dissociation constant (Kd) of 14.9nM at a density (Bmax) of 17.2 fmol/106 cells (Fig. 2A). We alsostudied the hUG binding parameters on NIH 3T3 cells. As withhuman trophoblasts, the NIH 3T3 cells expressed a proteinthat bound 125I-hUG with high affinity (Kd = 35.7 nM andBmax = 6.1 fmol/106 cells) (Fig. 2B).To further characterize this novel cell surface binding of

hUG, we performed affinity crosslinking studies with DSSusing 125I-hUG as the ligand in the absence or presence ofunlabeled hUG at both temperatures. The radioactive bandcorresponding to the putative hUG receptor on human tro-phoblast cell surface had an apparent molecular mass of =190kDa as resolved by SDS/PAGE under reducing conditions.This molecular weight was deduced after subtracting thecalculated molecular mass of homodimeric hUG (15.8 kDa)from the apparent molecular mass of the radioactivecrosslinked protein band. The radioactive crosslinked proteinbands are shown in lanes 1 of Fig. 3A (Left at 33°C; Right at

39°C). The high molecular weight radioactive protein band wasundetectable when 1 u-M cold hUG (Fig. 3A, lanes 2) was

included in the reaction mixture prior to crosslinking. Simi-larly, crosslinking studies with NIH 3T3 cells identified an

-190-kDa radioactive protein band corresponding to thehUG-binding protein (Fig. 3B, lane 2). This band was virtuallyabolished when 1 ,LM unlabeled hUG (Fig. 3B, lane 3) was

included in the incubation mixture for competition prior toaffinity crosslinking with DSS. As expected, in the absence ofDSS, no high molecular weight radioactive protein band was

visible in the autoradiogram (Fig. 3B, lane 1).Since many cell surface receptors are glycosylated proteins,

we sought to determine whether the hUG-binding protein isalso glycosylated. Thus, we performed deglycosylation studiesby pre- and posttreatment of the human trophoblast and NIH3T3 cells by N-glycosidase F and BSA-free O-glycosidase. Wefound that not only the apparent molecular mass of the125I-hUG-crosslinked protein band remained unaltered aftersuch treatment, but also the binding kinetics of 125I-hUG was

unaffected (data not shown). These results suggest that thehUG-binding protein is nonglycosylated and that posttransla-tional modification (i.e., glycosylation) of this protein is not a

prerequisite for binding the ligand.One of the most striking findings of our study was that while

affinity crosslinking of 125I-hUG to its binding protein was

readily detectable on trophoblast and NIH 3T3 cells, we couldnot detect any such binding protein using two different cho-riocarcinoma cell lines (Fig. 3C). Whether the absence of125I-hUG binding is due to a lack of expression of the bindingprotein on choriocarcinoma cell surface or is due to expressionB

0.18

0.12

0.6 \.

0 4 8 1

0 4 8 12Bound, fmol per 106 cells

FIG. 2. Scatchard plot of specific binding of 125I-hUG on human trophoblast (A) and NIH 3T3 (B) cells. 125I-hUG was incubated with eitherhuman trophoblasts or NIH 3T3 cells at 4°C for 2 hr in the absence or presence of different concentrations of unlabeled hUG. The data are fromthree experiments, and each point represents the mean of triplicate determinations.

. :, .c·O·· .~··· ··~~.: ,,· ,,'·

c,e,¢,, ~ %'* ...:.. "s e*': .~~~~~~~~~~~,?0 , .

,**~.A ***<:^'...

· ',* ?:- '·" · -''

4^ - t A °' 44?4q4..Q̂ ^ t 4.c

A. .- -, a

b..* ^ .A O -s

'c ^ ' ~". »ftt**-;«y"o t-Po "'

Y.

4F ? P ·Q .,

^ '! ."

' 44'4 , ' f ^

4444- 44 *4

^*:-^~z, (Q··9;4.4

IIU

Biochemistry: Kundu et al.

n

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996)

Ori kDa200.0-

116.4-

55.6-

26.6-14.3-

C 1 2 3 1 23kDa200.0-

116.4-

55.6-

2 3- Ori

- Ori

26.6-14.3-

FIG. 3. Affinity crosslinking of hUG-binding protein on humantrophoblast, NIH 3T3, and human choriocarcinoma cells. 125I-hUGwas incubated in the absence or presence of unlabeled hUG forbinding and then crosslinked with DSS. (A) Human trophoblasts.(Left) Cells grown at 33°C for proliferation. (Right) Cells grown at 33°Cand then transferred to 39°C. Lanes: 1, without unlabeled hUG, withDSS; 2, with unlabeled hUG and DSS. (B) NIH 3T3 cells. Lanes: 1,without DSS and unlabeled hUG; 2, with DSS, without unlabeledhUG; 3, with both DSS and unlabeled hUG. (C) Human choriocar-cinoma cells. (Left) JAR cells. (Right) JEG cells. Lanes: 1, without DSSand unlabeled hUG; 2, with DSS, without unlabeled hUG; 3, with bothDSS, and unlabeled hUG (NIH 3T3 and human choriocarcinoma celllines JAR and JEG were grown at 37°C). Note the absence of highmolecular weight radioactive protein band in C.

by these cells of a mutant protein that does not bind hUG isnot as yet clear.The uteroglobin family of proteins possess potent antiin-

flammatory and immunomodulatory properties (5-7). Humanuteroglobin gene expression is detectable in several organs(8-14, 17), and high levels of expression occur in the respira-tory and reproductive tracts of both sexes. Besides rabbits andhumans, this protein has been found in mice (33, 34), rats (15)and nonhuman primates (our unpublished results). The pres-ence of specific cell surface hUG-binding sites on humantrophoblasts as well as on NIH 3T3 cells and the dramaticantiinvasive effect of hUG on both of these cell types derivedfrom two different species of mammals suggest that thisproperty of hUG is not species specific. This is not totallyTable 1. Effect of recombinant hUG on extracellular matrix(Matrigel)-invasion of human trophoblast, human choriocarcinoma,and NIH 3T3 cells

Invasion,t %Cell type Treatment* of control

Human trophoblast None 100hUG 18Myoglobin 97

Human choriocarcinoma None 100hUG 100Myoglobin 100

NIH 3T3 None 100hUG 21Myoglobin 98

*Concentrations of hUG and myoglobin were 1 ,uM.tData represent the average of three experiments except for those withhuman trophoblasts, which were done in duplicate.

surprising as the UG gene is detectable in all mammalianspecies studied so far, and there is considerable structuralsimilarity between hUG and mouse UG (33, 34). The antiin-vasive effect ofhUG on human trophoblasts and NIH 3T3 cellsadds a new dimension to the list of multifaceted biologicaleffects of this protein. The potent antiinvasive effect of hUGon human trophoblasts and NIH 3T3 cells expressing thehUG-binding protein, and a lack of this inhibitory effect ontwo different human choriocarcinoma cell lines that do notbind hUG raise the possibility that the genes coding for hUGas well as its binding protein have important physiologicalfunctions in regulating cellular migration and invasion espe-cially during embryo implantation. Our results underscore thenecessity for further molecular characterization of the putativecell surface receptor of hUG and the mechanism of signaltransduction through this receptor-mediated pathway affect-ing cellular invasiveness.

Note. While this paper was in press we noted that Diaz Gonzalez andNieto (35) found a UG-binding protein with a molecular mass of 90kDa as determined by gel filtration. These authors have used reducedrabbit UG for their binding experiments. Our affinity crosslinkingexperiments using reduced hUG revealed that, in addition to the190-kDa protein reported in our manuscript, we could now detectanother UG-binding protein with a molecular mass of 49 kDa. Thisband was not detectable when nonreduced hUG was used forcrosslinking. Cloning of the cDNAs will delineate whether these twomolecular forms are subunits or two independent UG-binding pro-teins.

We thank Dr. Janice Y. Chou for a generous gift of the transformedhuman trophoblast and choriocarcinoma cell lines, valuable sugges-tions during the study, and critical review of the manuscript. The expertphotomicrography by Mr. Rick Dreyfuss and Ms. Shauna Everett ofthe Medical Arts and Photography Branch of National Institutes ofHealth and the reprinting of some of the color photographs by Mr. H.James Bahre of Fujix Electronic Imaging are gratefully acknowledged.1. Kliman, H. J. & Feinberg, R. F. (1990) Proc. Natl. Acad. Sci. USA

87, 3057-3061.2. Strickland, S. & Richards, W. G. (1992) Cell 71, 355-357.3. Krishnan, R. S. & Daniel, J. C., Jr. (1967) Science 158, 490-492.4. Beier, H. M. (1968) Biochim. Biophys. Acta 160, 289-291.5. Miele, L., Cordella-Miele, E. & Mukherjee, A. B. (1987) Endocr.

Rev. 8, 474-490.6. Mukherjee, A. B., Cordella-Miele, E., Kikukawa, T. & Miele, L.

(1988) Adv. Exp. Med. Biol. 231, 135-145.7. Miele, L., Cordella-Miele, L., Mantile, G., Peri, A. & Mukherjee,

A. B. (1994) J. Endocrinol. Invest. 8, 679-692.8. Peri, A., Cowan, B. D., Bhartiya, D., Miele, L., Nieman, L. K.,

Nwaeze, I. 0. & Mukherjee, A. B. (1994) DNA Cell Biol. 13,495-503.

9. Kleis-SanFrancisco, S., Hewetson, A. & Chilton, B. S. (1993)Mol. Endocrinol. 7, 214-223.

10. Dhanireddy, R., Kikukawa, T. & Mukherjee, A. B. (1988) Bio-chem. Biophys. Res. Commun. 152, 1447-1454.

11. Manyak, M., Kikukawa, T. & Mukherjee, A. B. (1988) J. Urol.140, 176-182.

12. Kikukawa, T., Cowan, B. D., Tejada, R. I. & Mukherjee, A. B.(1988) J. Clin. Endocrinol. Metab. 67, 315-321.

13. Singh, G., Singh, J., Katyal, S. L., Brown, W. E., Kramps, J. A.,Paradis, I. L., Dauber, J. H., Mcpherson, T. A. & Squeglia, N.(1988) J. Histochem. Cytochem. 36, 73-80.

14. Singh, G., Katyal, S. L., Brown, W. E., Phillips, S., Kennedy,A. L., Anthony, J. & Squeglia, N. (1988) Biochim. Biophys. Acta950, 329-337.

15. Singh, G., Katyal, S. L., Brown, W. E., Kennedy, A. L., Singh, U.& Wong-Chong, M. L. (1990) Biochim. Biophys. Acta 1039,348-355.

16. Mantile, G., Miele, L., Cordella-Miele, E., Singh, G., Katyal, S. L.& Mukherjee, A. B. (1993) J. Biol. Chem. 268, 20343-20351.

17. Peri, A., Cordella-Miele, E., Miele, L. & Mukherjee, A. B. (1993)J. Clin. Invest. 92, 2099-2109.

BA 330C 390C

kDa1 2 1 2

kDa200.0- *-.

116.4-

55.6-

26.6-

14.3-

2918 Biochemistry: Kundu et al.

LI

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

Biochemistry: Kundu et al.

18. Umland, T. C., Swaminathan, S., Furey, W., Singh, G., Pletcher,J. & Sax, M. (1992) J. Mol. Biol. 224, 441-448.

19. Matthews, J. H., Pattabiraman, N., Ward, K. B., Mantile, G.,Miele, L. & Mukherjee, A. B. (1994) Proteins Struct. Funct.Genet. 20, 191-196.

20. Levin, S. W., Butler, J. D., Schumacher, U. K., Wightman, P. D.& Mukherjee, A. B. (1986) Life Sci. 38, 1813-1819.

21. Miele, L., Cordella-Miele, E., Facchiano, A. & Mukherjee, A. B.(1988) Nature (London) 335, 726-730.

22. Miele, L., Cordella-Miele, E. & Mukherjee, A. B. (1990) J. Biol.Chem. 265, 6427-6437.

23. Facchiano, A., Cordella-Miele, E., Miele, L. & Mukherjee, A. B.(1991) Life Sci. 48, 453-464.

24. Waite, M. (1990) Adv. Exp. Med. Biol. 279, 1-22.25. Dennis, E. (1983) in The Enzymes, ed. Boyer, P. (Academic, New

York), 3rd Ed., vol. 16, pp. 307-353.26. Mukherjee, A. B., Miele, L. & Pattabiraman, N. (1994) Biochem.

Pharmacol. 48, 1-10.

Proc. Natl. Acad. Sci. USA 93 (1996) 2919

27. Vasanthakumar, A., Manjunath, R., Mukherjee, A. B., Warabi,H. & Schiffman, E. (1988) Biochem. Pharmacol. 37, 389-393.

28. Robinson, D. H., Kirk, K. L. & Benos, D. J. (1989) Mol. Cell.Endocrinol. 63, 227-234.

29. Peri, A., Dubin, N. H., Dhanireddy, R. & Mukherjee, A. B.(1995) J. Clin. Invest. 96, 343-353.

30. Lei, K. J., Gluzman, Y., Pan, C. J. & Chou, J. Y. (1992) Mol.Endocrinol. 6, 703-712.

31. Melchiorri, A., Carlone, S., Allanvena, G., Aresu, O., Parodi, S.,Aaronson, S. A. & Albini, A. (1990) Anticancer Res. 10, 37-44.

32. Tarentino, A. L., Gomez, C. M. & Plummer, T. H. Jr. (1985)Biochemistry 24, 4665-4671.

33. Ray, M. K., Magdaleno, S., O'Malley, B. W. & DeMayo, F. J.(1993) Biochem. Biophys. Res. Commun. 197, 163-171.

34. Singh, G., Katyal, S. L., Brown, W. E. & Kennedy, A. L. (1993)Exp. Lung Res. 19, 67-75.

35. Diaz Gonzalez, K. & Nieto, A. (1995) FEBS Lett. 361, 255-258.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021