De novo synthesis and secretion of a 36 kD protein by cells that form lupus inclusions in response to alpha interferon

(1)

De novo synthesis and secretion of a 36-kD

protein by cells that form lupus inclusions in

response to alpha-interferon.

S A Rich

J Clin Invest.

1995;

95(1)

:219-226.

https://doi.org/10.1172/JCI117643

.

In response to the pure recombinant human alpha-IFN, IFLrA, Raji and Daudi were the only

two cell lines among 19 human lymphoblastoid cell lines tested that formed the human

lupus inclusions (LI) to a high frequency. Raji, Daudi, and five other cell lines were

examined for protein changes that might accompany LI formation. Their selection was

based upon T or B origin, association with Epstein-Barr virus, and ability to form LI. A trace

protein of an estimated molecular mass of 36 kD (p36) and an isoelectric point of 5.6 was

detected on two-dimensional gels only of alpha-IFN-treated Raji and Daudi cells.

Gamma-IFN did not induce p36 or LI in any of these seven cell lines. In Daudi cells p36 and LI

formed simultaneously in response to IFLrA, and persisted until the alpha-IFN-induced

death of the culture. In Raji cells, p36 and LI appearance and disappearance coincided with

the addition and removal of alpha-IFN. Fractionation of Raji cells with nonionic-detergent

buffer placed p36 with the inclusions in the cytoplasmic supernatant. With detergent-free

buffer p36 and LI were distributed evenly between the nuclear and cytoplasmic fractions.

Pulse-chase experiments revealed that p36 was secreted. The de novo synthesis of p36

with alpha-IFN treatment was shown by labeling the cell proteins with [35S] methionine

before and after the addition of alpha-IFN. These results along […]

Research Article

(2)

De Novo

Synthesis and Secretion

of

a

36-kD Protein

by

Cells That Form

Lupus

Inclusions

in

Response

to

a-Interferon

Steven A. Rich

Laboratory ofPathology, Wadsworth Center forLaboratories andResearch,New York StateDepartment of Health, and SchoolofPublic

Health, State University ofNew York, The UniversityatAlbany, Albany,New York 12201-0509

Abstract

Inresponsetothepurerecombinant humana-IFN,IFLrA,

Raji and Daudiweretheonlytwocell linesamong19 human

lymphoblastoid cell lines tested that formed the human

lu-pusinclusions (LI) to a high frequency. Raji, Daudi,and

five other cell lineswereexamined forprotein changes that

might accompanyLIformation. Their selection wasbased

uponTorBorigin,association withEpstein-Barr virus,and

abilitytoformLI.A traceproteinofanestimated molecular

mass of 36 kD (p36) and an isoelectric point of 5.6 was

detectedontwo-dimensionalgels onlyofa-IFN-treatedRaji

and Daudi cells. Gamma-IFN didnot induce p36 or LIin

any of these seven cell lines. In Daudi cells p36 and LI formedsimultaneouslyinresponsetoIFLrA,andpersisted

until the a-IFN-induced death of the culture. InRaji cells, p36and LI appearance and disappearance coincided with the addition and removal of a-IFN. Fractionation ofRaji cells withnonionic-detergent bufferplaced p36with the

in-clusions in the cytoplasmic supernatant. With detergent-free bufferp36and LIweredistributed evenly between the

nuclear andcytoplasmicfractions.Pulse-chase experiments

revealed thatp36 was secreted. The de novo synthesis of

p36 with a-IFN treatmentwas shown by labelingthe cell

proteins with [35S]methioninebeforeand afterthe addition

of a-IFN. These resultsalong with previous results onthe denovosynthesisof LI in theendoplasmicreticulum(which

is involved in theprocessing and secretion of proteins)

sug-gestarole for LI in the synthesis and secretion of p36.(J.

Clin. Invest. 199595:219-226.) Key words: cell

differentia-tion*inclusion bodies *electrophoresis * lupus

erythemato-sus* acquired immunodeficiency syndrome.

Introduction

Humanlupusinclusions(LI)' (1)areabnormalcytoplasmic

structuresthat form in thecytoplasmofreticuloendothelial cells

AddresscorrespondencetoDr. Steven A.Rich, Wadsworth Centerfor

Laboratories andResearch,New YorkState Department of Health,

Em-pireStatePlaza, PO Box 509, Albany, NY 12201-0509. Phone:

518-474-6389;FAX: 518-474-7992.

Receivedfor publication 28 April 1994 andinrevised form 21July

1994.

1.Abbreviations usedinthispaper:IFLrA, human recombinanta-IFN

A; LI, lupusinclusion;OAS, oligoadenylate synthetase; p36, 36-kD

a-IFN-inducedprotein.

TheJournal of Clinical Investigation, Inc. Volume95, January 1995,219-226

of individuals with SLE (2) and AIDS (3). Virus-like

struc-tures, tubuloreticular inclusions and tubuloreticular structures areothernamesusedintheliterature to refer to LI. Such

inclu-sionsare notdetected in the cells ofhealthyindividuals. These

inclusions areinduced both in vitro andin vivo by type I(a and/; 1) butnottype1 (y)IFN(4).The inclusions are known to beproducts of normal cells abnormally stimulated with a-IFN asshownbytheirinductioninperipheral blood mononu-clear cells prepared from adult Red Cross blood donors (4),

andfrom umbilical cord bloods from routine deliveries (5). The formation ofLI in SLE and AIDS is the consequence of chronichighlevels of endogenous IFNin the circulationof

individuals with these diseases (6, 7). Persistently high levels

ofIFN contrast sharply with the typical 24-h burst of IFNthat

isassociated withavirus infection. The IFN in SLE and AIDS isan unusual a-IFN that is acid-labile like y-IFN, butthat is

neutralizedby antibodies to a-IFNs(6, 8). This unusual a-IFN

inthe seraof individuals with SLE and AIDS hasan

extraordi-nary abilitytoinduce the inclusions (8).

LIarereported to be a prognostic marker for disease pro-gression in individuals with AIDS (3, 9). Induction ofLI by SLE sera corresponds to the disease activities of the serum

donorsattheextremesof disease activity (8). Individuals with the greatestserumconcentrationsof IFN induce LI to the

great-est frequencies, and these patients have the greatest disease

activities. Individuals in remission donothave measurable

lev-els of IFN and theirseradonotinducetheinclusions.

Undetect-ableserumlevels ofIFN aretypical of healthy individuals (8). Thusfor both AIDS andSLEthe chronic presenceof IFN and formation ofLIis associated with diseaseactivity.

LIaredenovosynthesizedcomplexes of ribonucleoprotein and membrane with carbohydrate andno DNA(10). Theyform

inarestrictedregion of the endoplasmic reticulum that makes

contactwithadjacent regions of theouternuclearenvelope and

theGolgi apparatus (11). Theseresults suggest thatLImight

function in membrane biogenesis, the trafficking of proteinsto

the plasma membrane orto cytoplasmic vesicles, or the pro-cessing of proteins for secretion.

Little else is known about the origin, cell function, and

systemic significance of these inclusions because they have

neverbeenpurified and their molecular components have never been identified and characterized (12).An inabilitytoreadily purify these inclusions is consistent with their localization in

the endoplasmic reticulum, a membrane system not enclosed withinadoublemembrane andthereforenotabletobe isolated

in intact form (10).

Results from the current study show that an IFN-induced protein ofamolecular mass of 36-kD(p36)and anisoelectric point of 5.6 is associated withLI.This protein providesa new avenue for investigation into the possible significance ofthe

roles ofLI andIFNin the immune-based diseases ofSLE and

(3)

Methods

IFNpreparations and titers. The purified recombinant human IFNs IFLrA (Hoffmann-La Roche, Nutley, NJ) andgamma(Genentech Inc.,

South SanFrancisco, CA)wereassayed for antiviral activityonWISH

cells challengedwithvesicularstomatitis virus (Indiana strain) (13).

Their titers are reported relative to the National Institutes of Health

human IFN ca- and y- standards (G023-901-527 and Gg23-901-530,

respectively; kindly provided bytheAntiviral Substances Program of theNational Institutes ofAllergy and InfectiousDiseases). 1U/ml of

IFN provides50%protectionof cellmonolayers.

Celllines,culture and IFN induction. 19 humanlymphoblastoid cell

lines(seeTableI)weregrowninsuspension in RPMI 1640 medium/

10% fetal calfserum(GIBCO BRL,Gaithersburg,MD)/100 U/ml of penicillin and 100/ig/ml ofstreptomycin (14). Tissueculture flasks (Bellco Glass, Inc., Vineland,NJ)werekeptinahumidity-controlled incubator(Wedco,SilverSprings,MD)witha5% CO2 atmosphereat 370C. The cellswere maintained inexponential growth andat <5%

nonviability by resuspension every 3 or 4 d in fresh medium. Cell

densitiesweredetermined induplicateonamodel F cytometer(Coulter

Cytometry, Hialeah, FL), and the percentage of nonviable cells was

determinedbytrypan blue exclusion. Cellsweretaken fromculturesat

indicated times of 100U/mlofIFLrAory-IFNtreatmentforanalysis ofLI, p36,andoligo-isoadenylatesynthetase.

Epstein-Barrnuclearantigenassay. All 19 of the human

lympho-blastoidcell lineswereevaluated forEpstein-Barrnuclearantigen (15).

Multi-well slideswere_preparedfrom PBS-washed cells thatwerefixed

for5 mininice-cold methanol andacetone( 1:1, vol/vol),and air-dried.

Epstein-Barr positiveandnegativehumansera(virus testingprogram,

Wadsworth Center for Laboratories and Research) were decomple-mented and added toduplicatewells of cells. The slideswereprocessed

as described (15), mounted in pH 9 mounting fluid, and read on a

fluorescencelightmicroscopeusingaX40dry objectiveand X1O

eye-pieces.

[35S]Methionine radiolabeling ofcell and secretedproteins. Cells were washed in methionine-free RPMI 1640 media (GIBCO BRL)/

10%dialyzedFCS. 20 x 106 cellswereincubated for 40minat37°C in8 mlofmediumsupplementedwith 25_ACi/mlof[35S}methionine

(ICNBiomedicalsInc., Irvine, CA).For the detection of secreted

pro-teins, half of these radiolabeled cellswerewashed andresuspendedin 4ml of RPMI1640/0.75%FCS,incubated for2h,and thesupernatant

was_centrifugedat100,000gfor 60min (type50rotor,Beckman Instrs.

Inc.,Fullerton,CA)beforeprocessingfortwo-dimensional_{gel analysis.}

Forthedetection of cellproteins,theremainingcellswerewashed and

suspended in 2 ml of RPMI 1640/10% FCS, incubated for 5 min, then washed andsuspendedin 2 ml of serum-free RPMI 1640before

processingfortwo-dimensionalgelanalysis.Incorporated radioactivity

was determined from portions ofthe two-dimensional gel samples. IFLrAandy-IFNwereaddedtothecorrespondingwash and incubation

mediaas_required.

Cellfractionation. 50 x 106IFLrA-inducedRajicellswere

resus-pendedin4°C reticulocytestandard buffer/0.ImM phenylmethylsulfo-nylfluoridewith and without0.1%Tween 40 (10, 16). After 10min

onice the cellswere_gentlyvortexed untillysed. Nucleiwerepelleted at1500gfor 10min.LIwerepelletedfrom the nuclearsupernatantby

a15-min centrifugationat8,000g. Nuclei and LIpelletswereoverlaid withglutaraldehydefor electronmicroscopy,andprecipitatedwith TCA for two-dimensional_{gel analysis.}

Electronmicroscopeenumerationof LU.Cellsampleswerefixedin 3%glutaraldehyde,postfixedin osmiumtetroxide, dehydratedingraded ethanol, and embedded in Epon (1). Thin (0.25-jim) sections were stained with uranyl acetate and lead citrate, and examined on a 301

electronmicroscope(Philips Electronic Instrs.Co.,Mahwah, NJ). LI frequencyestimates arebasedon abinomialmodel(1).Theratio of

themeandiameter of inclusions to that of whole cells is theprobability (d) ofobserving an LI in arandom thin section ofa_given cell that containsanLI.Factoringin theprobabilitythatanLI exists in the cell

Table I. LI FormationinHumanTandBLymphoblastoid

Cell Lines

Cell line LIfirquency Cellorigin EBNAstatust

Daudi 15 B +

Raji 12 B +

Raji-R§

<2 B +

U698 <2 B

NC 37 <2 B +

Jiyoye <2 B +

EB-3 <2 B +

Namalwa <2 B +

SB <2 B +

HSB2 <2 T

-MOLT <2 T

-CEM 0 T

-BJAB11 0 B _

BJAB/HRIKII

0 B +

BJAB/B95-811 0 B +

RAMOS1 0 B

AWRAMOS1 0 B +

LUKII** 0 B

8866.7** 0 B

*_{Percentage of}₄₀₀_independent_{cell sections examined that contained}

anLI(seeEM enumeration ofLI; Methods).Cellswerecultured for 72 h in the presence of 100 U/ml of IFLrA. t The measured presence

(+)orabsence(-)ofEpstein-BarrVirus NuclearAntigenwas deter-minedbythecomplement-fixationfluorescence assay(seeEBNAAssay,

Methods). §_Raji-Rwasderived fromRajicellsbycontinuousgrowth in the presence ofbromodeoxyuridinetoselect cells thatwereresistant tothe toxic effects ofbromodeoxyuridine. 11BJAB/HRlK and BJAB/ B95-8werederived from theEBV-negativecellline BJABby transfor-mation in vitro with the Epstein-Barr viruses HRlK and B95-8, respec-tively. I_{AW RAMOS} _was_{derived from the EBV-negative cell line}

RAMOSbyinfection in culture with theEpstein-BarrvirusAW. ** LUK II and 8866.7 produceda-IFNconstitutively.

(P) gives aprobability of observing LIinacellpopulation (P*), or

P* = _dP._For_each_{sample 400 independent}_cell _sections_were

exam-ined. Thin sectionsseparated bygreater than the

20-/im

diameter of the cells were used to approximate random sampling. A 10% frequency (P*) (40outof 400independentcellsections)is consistent withone

2.0-jim diameter LI (P= 1) per cell of a20-,umdiameter (d = 2.0 jim/20tim= 10%). The corresponding 95% confidence interval of LI

frequenciesfor400 independent cell sections is 7-13%(1).

Two-dimensional polyacrylamide gel electrophoresis. Aliquots of 107 cells (1 mg ofprotein) werewashedin serum-free RPMI 1640, precipitatedwith 10% TCA (wt/vol), and extracted twice withethyl ether, dried, and processed for isoelectric focusing (17). Two-dimen-sional polyacrylamide gel protein maps wereobtained (17, 18) with 200

_til

of samplerunon3-mm inner diameter and 12-cm longfocusing cylinders. Cylinderswerefocusedat400 V for 16 h, and thenat 800 V for 1 h inaPolyanalyst gelelectrophoresis apparatus (HAAKE,Inc.,

Paramus, NJ) using 0.1 M phosphoric acid (anode) and 0.1 M sodium

hydroxide(cathode). Equilibrated cylinders (5 ml of 3% (wt/vol)SDS,

6.2% (wt/vol)/3-mercaptoethanol, 62.5 mM Tris, pH 6.8 for 15 min

(4)

Figure 1. Electron micrographs of lupus inclusions (LI) in thin sections(0.1 ym) of (a)Daudi and (b) Raji cells that were cultured for 72 h with 100U/ml ofIFLrA. N,nucleus;L,lipid granule. LI are seen to be complexes of membrane with tubular structures of 20-28 nm diameter. The sections were stained with uranyl acetate and lead citrate, and examined in a Philips 301 electron microscope. Bar= 1

_ym.

cylinders equilibrated in 1 ml ofwater. The vertical slab apparatus (SE500; Hoefer Sci. Instrs., San Francisco, CA) was run at 100Cand

110 Vfor 30min,followed by 220 V until the bromphenol blue dye marker reached the bottom of the slab (- 3 h). Slab gels were stained with either Coomassie brilliant blue or silver (Bio Rad Laboratories, Richmond, CA). For fluorography slab gels were fixed for 45 min in 50%methanol/ 10% aceticacid,soaked in water for 30min,and equili-bratedin 1 Msodium salicylate for an hour (19), before drying and processing (20).

Oligoadenylate synthetase (OAS)assay. Cellpelletswashed inPBS were overlaid with 500

Ml

of Triton X-100 lysis buffer (0.5% wt/vol Triton X-100, 20 mM Hepes, 120 mM KC1, 5 mM magnesium acetate, 7mM,6-mercaptoethanol, 10% wt/vol glycerol) and vortexed (5, 21). Supernatant (100

41)

after 2 min of centrifugation at 15,600 g was

adsorbed topoly(rI) (rC)-agarose beads (Pharmacia LKB Biotechnol-ogy Inc., Piscataway, NJ). The picomoles of ATP incorporated into isoadenylate oligomers were calculated from the cpm in the eluted prod-uctsfrom the alumina column (Sigma Chemical Co., St. Louis, MO) (5, 21) and the cpm of thea[32P]ATP(Amersham, Arlington Heights, IL) added to the adsorbed agarose beads (5).

Results

IFLrA induction of lupus inclusions in human lymphoblastoid cells. 19 human lymphoblastoid cell lines were evaluated for

theirabilitytoformlupus inclusions inresponse tothe purified recombinant leukocyteIFN, IFLrA(TableI). BeforeIFN treat-ment no inclusions were seen in anyof thecelllines. With a-IFN treatment two of the celllines, Daudi (Fig. 1 a) andRaji (Fig. 1 b), formed inclusionsat acell frequency of> 10% of

cell sections, nine at afrequency <2%, and in the remaining eight lines no inclusions were detected. LUK II and 8866.7 constitutively producea-IFN. Neitherofthesecell linesformed

LIeither beforeorafter the addition of IFLrA. Both Daudi and Raji cells formedLIwhengrownin media conditioned by these cells, which showedthatthe IFN theyproduced had the

biologi-cal activity to induceLI. LIformation was notrestricted to B orT cellorigin, orinfection with Epstein-Barrvirus (Table I,

EBNAstatus).

Two-dimensional protein gel analyses of IFN-induced cells. ThehighLIfrequencies in Daudi and Raji cells suggested that LI-associated IFLrA-inducedproteins would also be elevated. Comparison of two-dimensional protein gels revealed onlyone

IFLrA-induced protein thatwascommon toboth IFLrA-treated cells (Fig. 2, aand c;Daudi and Raji, respectively) that was not present inthe untreated cells (Fig. 2, b andd; Daudi and

Raji, respectively). This protein, termedp36,has an estimated

size of 36 kD andan isoelectric point of 5.6.

Five additional cell lines (Table II) chosen from those listed

inTableI werealsoexamined bytwo-dimensional gels for p36 protein. Selection of these cell lineswasbasedontheirB (SB, RAMOS, and AW RAMOS) and T (HSB2 and MOLT) cell origins, their association with Epstein-Barr virus (SB and AW

RAMOS) and their formation of LI, although at thelow fre-quencies of<2%(SB,HSB2, and MOLT). However, p36was not detected in any of these other cell lines either before or

afterIFN treatmentwith 100U/ml ofIFLrA.

Gamma-IFN did not induce p36 orLI in any of the cell linesexamined, including Daudi and Raji cells. OAS activation indicated thatall seven of these cell lines were responsive to

botha- and y-IFNs. Forall of these cell linesOAS activation

was lesswith y-IFN thanwithIFLrA. Daudi and Raji didnot

always show activation with y-IFN, and when activation

oc-curred, itwas always < twofold incontrast with a >twofold andup tofivefold (RAMOS) activationinthe other five lines. OAS activation withIFLrA was shown to betransient, rising

rapidly by 18-24 h and falling to almost untreated levels by

144 h.

TimecourseofIFLrAinductionof lupus inclusions and p36 protein inDaudiand Raji cells. In Daudi cellsa-IFN induced

(5)

a

94-

67-

43-

30-0-%

L.g

-(A

I1

Am.

..

A_

e^...

:.

:g-~iM~

.:."''.'.,,,~~~~~~',4

*!

.4O A= sa

20.!-144a

94-

67-

43-

30-

20.1-

144-4 .0 l.

4.0

*~~~~~~~~~3

* *-4h

5.0

5.5

6.0 7u

Isoelectric pH

value

Figure2.Two-dimensionalpolyacrylamide gel electrophoresis mapsofDaudi(ab)andRaji (c,d)cells after(a,c)andbefore(bd) culturing for

72 hwith 100 U/ml ofIFLrA. p36wasdetectedonlyafter IFLrA treatment (ac). Each sample consistsof 400

_jig

of total cellprotein.Molecular massesandisoelectricpointsarenoted on the vertical and horizontalaxes,respectively.Silverstainingwasusedbecause it is moresensitive than Coomassie brilliantblue,and itpreferentially enhances p36 relativetothesurrounding proteins.

todieby 48 h, which increasedto100%by 120 h. Both the LI

andp36 proteinwereapparentthroughoutthe120 h timeperiod

ofcontinued presence of a-IFN. In Raji cells

a-FEN

induced

thep36 proteinby 12 h,and the LIby24 h (Fig. 3 b). They

werepresentthroughoutthe 168 h of a-IFNtreatment.Unlike Daudicells, Raji cellgrowthandviability wereunaffectedby

IFLrA.By48hafter the removal of a-IFN the p36 proteinwas nolonger detected, and the inclusionswerereducedtothe very

lowfrequencyof 1.25%.Theessentially sametimecoursesof

appearance ofp36andLIupona-TFNstimulationof Daudi and

Raji cells,anddisappearanceupon removal of a-IFN withRaji

cells, provide additional evidence for the association ofp36

with LI.

Localizationof lupus inclusions and thep36proteinincell

fractions. Purified nuclei from IFLrA-induced Raji cells were

prepared with nonionic detergent buffer (Fig.4a). TheLIfrom

these cellswereidentified only in thecytoplasmic supernatant

(Fig. 4b). Fractionation in the absence of nonionic detergents

produced nuclei that contained alargeamountofcytoplasmic

membrane andLI(Fig. 4c). TheLI were also located in the

corresponding cytoplasmic fraction from these nuclei (Fig. 4

d). Two-dimensional gelsshowedthe presence ofp36 protein

in theonenuclear(Fig.4c)andtwocytoplasmic (Fig.4b and d) fractions that contained LI. No p36 was apparent in the nuclearpellet withoutLIthatwas prepared with the nonionic

detergent (Fig.4a).

Secretionofp36protein. Two-dimensionalgelsof radiola-beled proteins released into the media by Raji cells revealed

thatp36protein was secretedby theIFLrA-treatedcells

(Fig.

5 a), but notbytheuntreated cells (Fig. 5 b). Specificity for

theproteins beingsecreted ( 16.1% ofincorporated

[3H]-amino

acids)wasshownby comparisonwith

[3H]thymidine

and[3H]

-.%A. _M _-.

-IA-:: A ins NOWh ^E~~

(6)

Table II. Formationofp36 inSelectedCell Lines

p36 Protein* Oligoadenylate synthetaset Cell lines IFLrA y 0 IFLrA y 0 Daudi + - - 6177 1428 888 Raji + - - 4157 1252 721

SB - - - 2147 1472 375

HSB2 - - - 1065 185 90

MOLT - - - 1052 456 233

RAMOS - - - 2378 1023 203 AWRAMOS - - - 1687 843 188

*_{The presence (+) or absence (-) of}_p36_was_{determined on}

silver-stainedtwo-dimensional gels run with 400,gof each cell line before (0),and afterculturingfor 72 h with 100 U/mlof IFLrA(e.g.,Fig.2, aand b for Daudi andRaji, respectively)ory-LFN. tOligoadenylate synthetase assay values are reported as pmolofATPhydrolyzedper 1 X 105 cells(see Oligoadenylate synthetase assay, Methods) before(0)

andafter culturing for 72 h with 100 U/ml of IFLrA or y-IFN.

uridine (2.6 and 1.6%, respectively). No differences were ob-tained for theamountof secreted radiolabel between the

IFLrA-induced and untreated cells.

The chase mediumwasalso examined ontwo-dimensional gels from the other six cell lines listed in Table II. Daudi, like Raji, secreted p36 in chase medium only after IFLrA treatment. None of the other five cell lines showed the presence of p36 either withorwithout IFLrA treatment.

De novosynthesis ofp36 protein. Raji cells were radiola-beled with [35S] methionine for 18h(onepopulation doubling), chasedfor 6 h, and cultured for an additional 18 h in the

pres-enceof 100 U/mlof IFLrA. However, p36 was not detected in

an autoradiograph of these cells (Fig. 6 a), which indicated

thatp36wasnotformed fromproteinssynthesized before IFLrA addition. Yetp36was detected in autoradiographs of aliquots of these same cells whenpulsed atthe times of2 (Fig. 6 b) and 18 (Fig. 6 c) h after the addition ofIFLrA. This showed that thesecellswereactively synthesizing p36atthis time. Thus p36 was synthesized de novo in response to IFLrA, and not frommodificationstoproteins that were synthesized before the addition ofIFLrA.

Discussion

This study demonstrated that in Raji and Daudi cells a-IFN

induces LI inassociation witha36-kDproteinwith an isoelec-tricpoint of 5.6. Only these2celllines, among 19tested(Table I), formedLI tohighfrequencies and p36todetectable levels (Fig. 1, Table H). The association of p36 proteinwith LI was

further demonstrated by cellline responses to a- and y-IFNs

(Table II), by timecourse studies withthe addition and removal

of a-IFNtoRaji and Daudi cell cultures (Fig. 3),andbytheir

codistribution in nuclearandcytoplasmicfractions ofRajicells

(Fig. 4). Raji cellswerealsoshowntosynthesizep36denovo

(Fig. 6), like LI (10).

p36was notdetected in cell lines with low LIfrequencies. Thiswasprobably duetothetwo-dimensionalgel assay's rela-tively low sensitivity.LIfrequency is the ratio of theLI tocell diameter (1).AnLIfrequencyof 10% describes

2.0-,um

LIin 20-,um cells (2.0

_tm/20

um), and LI of 0.4 um in the same

diameter cell gives a 2% frequency (0.4 ,.sm/20 ,um).

Two-dimensionalgelsmeasurethe relative concentration of cell pro-teins, a volumetric relationship. Assuming that p36 is an LI

component, its tracepresence inRaji and Daudi cells (at 10%

orgreater LIfrequency) indicates thatat2%LI orless(Table

I)p36 wouldnotbe detected[1/125th theamount(0.4,um/2.0

Mm)3].

This analysis of other cell lines that didn't formLI above

a2% frequency concurred with nop36 detected in Raji cells with a< 2%frequency ofLI at48haftertheremoval ofIFLrA

(Fig. 3 b).Incontrast, the addition of IFLrA to both Raji and

Daudi cellsinduced p36 before the formation ofLI (Fig. 4 a at 6 h, and Fig. 4 b at 12 h). The subsequent exponential increase in LIfrequency suggested the accumulation of a critical

amountof LI precursors before the formation of the uniqueLI

structure, andthatp36 might be suchaprecursor.

LI are knownto be disparate from an antiviral state since they do not form in WISH and MDBK cells witha- and y-IFNtreatment(22). P36 differs frommostotherreported

IFN-induced proteins (23, 24) since it is not induced by y-IFN (Table II, 4), and it is undetectable without a-IFN treatment (Figs. 2 b and d, 5 b, and 6 a).

Rapid secretion of p36 could explain its absence from cell lineswithoutLI. Incells with LI, p36 might indicate an inhibi-tion of this rapid secretion. However, p36 was not secreted

18

a

o0--

Figure3. Timecourseof IFLrA

sC a (100U/mi) inductionofp136(4-)

15 o8o\ and LI (o) in Daudi (a) andRaji

>-

12 (b) cells.p36 is indicatedas

detect-O ableorabsent fromtwo-dimensional

Z 0 6

Uj

/ ol

\gels

of

corresponding

time

point

0X

O / 4 I t samples. LI frequency was the

per-E 6 / t \ _centageof 400

_independent

cell

sec-3 3 / 36kD protein detected ₀ 2 36 kD protein detected tions examined that contained an LI (see EM enumeration of LI; Meth-ods). Incontrast tothecomplete

0 612 24 48 72 96 120 224 48 72 96 120 144 168 -24 -48 -2 -6 IFLrAkillingoftheDaudicellculture 4 +IFLrA * * +IFLrA < -IFLrA_p- _{by 120 h (% nonviable cells of 9, 54,} H H 86, and 100 at 48, 72, 96, and 120 h, respectively), Rajicells were

unaf-fected.Thetimecourseofdisappearanceofp36 and LI began at 168 h with the removal of IFLrA. The asterisk in (a) denotes the detection, but

(7)

w'..

I ..2

Figure4. Fractionation of p36with LI. Nuclear (a,c) and cytoplasmic (b,d) fractions of 72 h IFLrA-induced (100 U/ml)Rajicells were prepared with the low ionic strength buffer reticulocyte standard buffer in the presence (ab) and absence (c,d) of the nonionic detergent Tween 40 (0.1%, wt/vol). N, nucleus. LI distribution was estimated from the average number of LI per grid square in the nuclear, and cytoplasmic (8,000 g for 20

min),pellets. With Tween40 allLI wereinthe cytoplasmic fraction (b; e.g.,no LI weredetected in the corresponding nuclear pellet, a). Without Tween40 LI distribution was approximately equal between the nuclear (c) and cytoplasmic (d) fractions. p36 was identified only in the three fractions in which the LI were present (b-d). Electron micrographsareof thin sections (0.1 [m)that werestained with uranyl acetate and lead citrate,and examined inaPhilips 301 electron microscope. Bar = 1

_ytm.

eitherinthe absenceor presence of a-IFNby cell lineswithout The a-IFN-activated intracellular pathway consists of the

LI, and it was among Raji and Daudi cell-secreted proteins nuclear translocation of threecytoplasmic proteins (p84, p91,

upon a-IFN treatment (Fig. 5). The < 3% ofthymidine and andp113) aftertheirphosphorylation atspecific tyrosine resi-uridinedetected in aliquots ofthesesamples in contrast to the dues, their combining with a nuclear resident protein (p48),

16% of3H amino acids indicates specificity forprotein secre- andtranscriptional activation of a-IFN-specificgenes(25). Cell

tion, and shows that LI did not inhibit it. LI formation in a activation by y-IFN involvesthephosphorylation of the 91-kD restricted endoplasmicreticulumregionthat makes contact with cytoplasmicresidentprotein,whichformsadimer,translocates

the adjacentouter nuclear envelope and Golgi apparatus (11) tothenucleus,and activatesy-IFN-specificgenes(26).

(8)

LA

I

r

-Afi 4w wamF

_m

0 ., g

r

0 OF~~

I, .. I

5.0 5.5 6.0

b

7.0

4.0

5.0:

5.5 6.0 7.0

Isoelectric

pH

Value

Figure 5. Two-dimensionalgel autoradiographsofproteinssecretedby Rajicells treated with 100 U/mlofIFLrA for 72 h(a),andnottreated withIFLrA(b). p36 was secreted only by Raji cells afterLFLrAtreatment(a), but not in the proteins secreted by untreatedRajicells(b).Inboth

samples the amounts of3H-aminoacids(16.1% ofthe totalincorporated duringthepulse), or[3H]thymidine(2.6%)or [3H]uridine(1.6%)

measured in thesecretorymedium were the same. Theapproximatelyeightfoldgreater release of proteinoverDNA and RNAdemonstratesa

selective secretionof protein,andp36 bya-IFN-treated cells.

LIandp36 couldberelevanttoa-IFN'simmunemodulatory

effects (27-29). Two-dimensionalgelsofa-IFN-induced

hu-manlymphoblastoid (30,31 )orperipheralbloodmononuclear

(32, 33) cells revealaproteinwithp36's approximatemolecular mass and isoelectric point. LI, y-IFN, cell fractionation and

secretion of proteins were not investigated in these studies. Daudi but notRaji cells wereincluded in the lymphoblastoid

studies (30, 31), and therefore it appeared that the

a-IFN-inducedprotein changes wereassociatedwith Daudicell's se-veregrowthinhibitionand death (Fig. 3a).

97

~67-'A ₄₃ _'

30 ' @ ' k

(U-With silver staining p36 is visible in whole cell samples (Fig.2aandc), and withCoomassiebrilliant blueupon

enrich-ment.Thisindicatedthatp36was acandidateproteinfor

purifi-cation andfurther characterization (17, 34, 35). Tryptic peptide microsequence of purified p36 did not share any significant sequenceidentitywithproteinslisted in theSWISS-PROT (Ge-neva,Switzerland)database, suggestingthatit isa newprotein

(Rich,S.A., unpublished results). Anti-peptide antiserum

pre-pared against p36 synthetic peptides has detected p36protein

in Western blots ofRajicells only after IFLrA induction, and

IsoelectricpHValue

Figure6. Two-dimensionalgel autoradiographs ofRajicellproteins. Rajicellswereradiolabeled with [35S ]methionine for 18 h(one population doubling),chased for 6h,and cultured for 18 h with 100U/mlof IFLrA(a). Aliquotsofthesesamecellswereradiolabeled for 40minafter

culturingfor 2(b)and 18(c)hwith IFLrA.Togethertheseresults showed thatp36wasnotformed fromproteinssynthesized (radiolabeled)

before the addition of IFLrA(a),and itssynthesisafter 2(b)and 18(c)h of IFLrAtreatmentwasadenovoresponsetoIFLrA. a

94

67_ht

_. 43 0

a.-5 30

U

0

20.1-12OA

4.0

I.

(9)

at anestimated>400-foldinduction level. Confirmation ofthe

microsequence results will require amplification, cloning, and sequencing of p36 mRNA cDNA. In situ immunoelectron mi-croscopy with a suitable antibody preparation should reveal the cellular location of p36 and its relationship with LI.

Bcellproduction and secretion of p36 suggests the possibil-ity for interleukin activpossibil-ity. Another possibilpossibil-ity is that secreted p36 induces autoantibodies. Thus a-IFN in individuals with SLEand AIDS could stimulate LI formation and p36 production andsecretion. Also, p36 could modulate the immune system by interleukin or autoantigen activities. Further characterization of p36 and LI should provide additional understanding of a-IFN's roles in SLE and AIDS immunopathology.

Acknowledgments

The expert technical assistance of William E. Gibbons is greatly ac-knowledged. Iamindebted to Drs. Himangshu and Mahuya Bose for reading the manuscript, to Dr. Sidney Pestka (Robert Wood Johnson Medical School) for providing the IFLrA, to Mr. Scott Geromanos

(Sloan-Kettering Institute for Cancer Research, Rockefeller Research Laboratories) for the microsequencing of purified p36, to Dr. Lee H.

Kronenberg (Lee Biomolecular Research Laboratories\ Inc.) for the LUKII and 8866.7 celllines, and to Dr. Julius Youngner (School of

Medicine, UniversityofPittsburgh) forthe RAMOS and AW RAMOS

celllines.

Thiswork wassupportedbyagrant from theNational Institutes of Health (AR41619toS. A. Rich).

References

1.Rich,S. A. 1981. Humanlupusinclusions and interferon. Science(Wash.

DC). 213:772-775.

2.Gyorkey,F., K.-W. Min, J. G.Sinkovics,and P.Gyorkey.1969.Systemic

lupuserythematosus andmyxovirus (letter).N.Engl. J. Med. 280:333. 3. Orenstein,J. M.,G. L. Simon,C. M. Kessler,and R. S. Schulof. 1985. Ultrastructural markers incirculating lymphocytesofsubjectsatriskfor AIDS. Am.J.Clin. Pathol. 84:603-609.

4.Rich,S.A.,T. R.Owens,L. E.Bartholomew,and J. U.Gutterman. 1983. Immuneinterferon does not stimulate formation ofalpha and betainterferon induced humanlupus-typeinclusions(letter).Lancet.i:127-128.

5. Rich,S. A.,and W. E. Gibbons. 1990. Humanlupus-typeinclusions in umbilical cord blood mononuclearcells. Arthritis Rheum. 33:1420-1425.

6. Preble,0.T.,R. J.Black, R. M.Friedman, J.H.Klippel,and J. Vilcek. 1982. Systemic lupuserythematosus: presence in humanserumofanunusual acid-labileleukocyteinterferon. Science(Wash. DC). 216:429-431.

7.DeStefano, E.,R.M.Friedman,A.E. Friedman-Kien,J.J. Goedert,D. Henriksen, 0.T. Preble,J. A. Sonnabend,and J. J.Vilcek. 1982. Acid-labile human leukocyte interferon in homosexual men with Kaposi's sarcomaand

lymphadenopathy. J.Infect.Dis. 146:451-459.

8.Rich,S.A.,T. R.Owens,M.C.Anzola,and L. E. Bartholomew. 1986. Induction oflupusinclusionsbyserafrompatientswithsystemic lupus erythema-tosus.Arthritis Rheum. 29:501-507.

9.Feremans,W. W., K.Huygen,R.Menu,C. M. Farber,J. P. DeCaluwe,

J. P.VanVooren, L. Marcelis,L.Andre, M. Brasseur,H. Bondueetal. 1988.

Fiftycases of humanimmunodeficiency virus infection: immunoultrastructural

studyofcirculatinglymphocytes.J. Clin. Pathol. 41:62-71.

10.Rich,S.A.,and A. A.Reilly. 1989. Evidence that the interferon-induced Daudi cell humanlupusinclusionsaredenovosynthesized complexesof

ribo-nucleoproteinand membrane. J.Biol. Chem. 264:3529-3537.

11.Rich,S. A., M.Marko,and W. E. Gibbons. 1992. Localization of

inter-feron-inducedlupusinclusions demonstrated by computer image reconstruction ofmonensin-treated Daudi cells. J. Struct. Biol. 108:25-34.

12.Rich,S.A., and W. E. Gibbons. 1990. Nature and origin of lupus-type inclusions (letter). Hum. Pathol. 21:1191-1192.

13. Rubinstein, S., P. C. Familletti, and S. Pestka. 1981. Convenient assay for interferons. J. Virol. 37:755-758.

14. Rich, S. A. 1981. Induction of lupus inclusions by BUdR is mediated by ahuman lymphoid cell factor. Arthritis Rheum. 24:1444-1447.

15.Reedman, B. M., and G. Klein. 1973. Cellular localization of an Epstein-Barr virus (EBV)-associated complement-fixing antigen in producer and non-producer lymphoblastoid cell lines. Int. J. Cancer. 11:499-520.

16.Penman,S.1969. Preparationof purified nuclei and nucleoli from mamma-liancells. In Fundamental Techniques in Virology. K. Habel, and N. P. Salzman, editors. Academic Press, New York. 35-48.

17.Rich, S. A., and W. E. Gibbons. 1989. ROTOFOR cell isoelectric point fractionation ofdetergent solubilized protein cell samples. Technique(Phila.).

1:196-203.

18. O'Farrell, P. H. 1975. Highresolution two-dimensionalelectrophoresisof

proteins. J.Biol. Chem. 250:4007-4021.

19. Chamberlain, J. P. 1979. Fluorographic detection of radioactivity in

poly-acrylamide gelswiththe water-soluble fluor, sodium salicylate. Anal. Biochem. 98:132-135.

20. Bonner, W. M., and R. A. Laskey. 1974. A film detection method for tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46:83-88.

21. Merlin, G., M. Revel, and D. Wallach. 1981. The interferon-induced enzyme oligo-isoadenylate synthetase: rapid determination of its in vitro products. Anal.Biochem. 110:190-196.

22. Rich,S. A., and T. R. Owens. 1984. Purified recombinant human leukocyte interferons IFLrA and IFLrD induce human lupus inclusions in Raji and Daudi cells. J.Interferon Res. 4:335-345.

23.Beresini, M. H., M. J. Lempert, and L. B. Epstein. 1988.Overlapping

polypeptide induction in human fibroblasts in response to treatment with inter-feron-alpha,interferon-gamma,interleukin1-alpha,interleukin1-beta,and tumor necrosisfactor. J. Immunol. 140:485-493.

24.Sen,G. C., and R. M. Ransohoff. 1993. Interferon-induced antiviral actions andtheir regulation. In Advances in Virus Research. K. Maramorosch, F. A. Murphy, and A. Shatkin, editors. Academic Press, San Diego. 57-102.

25.Shuai, K. 1994. Interferon-activated signal transduction to the nucleus. Curr.Opin. Cell Biol. 6:253-259.

26.Shuai, K., G. R.Stark,I. M. Kerr, and J. E. Darnell, Jr. 1993. Asingle

phosphotyrosine residue of STAT91 required for gene activationby interferon-gamma. Science(Wash.DC)261:1744-1746.

27.Holan, V., S. Nakamura, and J. Minowada.1992.Inhibitory versus stimula-tory effects of natural humaninterferon-alpha onproliferation oflymphocyte

subpopulations. Immunology. 75:176-181.

28.Bloom, B. R. 1980. Interferon and the immune system. Nature(Lond.).

284:593-596.

29.Herberman,R. B., J. R.Ortaldo,and G. D. Bonnard. 1979.Augmentation

by interferon of human natural andantibody-dependentcell-mediatedcytotoxicity.

Nature(Lond.). 277:221-223.

30. Guardini, M. A., M. P.Schoenberg, R. B. Naso, B. A.Martin, J. U.

Gutterman,and J.Guevara,Jr.1984. Induction ofproteinsin human

lymphoblas-toid cells by recombinant alpha interferon. J. Interferon Res. 4:67-79. 31.Leanderson,T., S.Sundstrom,I.-L.Martensson,T.Ny,and E.Lundgren.

1982. Interferon-specific effects on protein synthesis inP3HR-1 cells. EMBO (Eur. Mol. Biol. Organ.) J. 1:1505-1511.

32.Cooper,H. L. 1982.Effect ofbacterially producedinterferon-a2on

synthe-sisofspecific peptidesin humanperipherallymphocytes.FEBS(Fed.Eur.

Bio-chem.Soc.)Lett. 140:109-112.

33. Gustafsson, A., S. Sundstrom, T. Ny,and E. Lundgren. 1982. Rapid

induction of sevenproteinsin humanlymphocytesbyinterferon;correlationto

natural killer cellactivity.J. Immunol. 129:1952-1959.

34.Matsudaira,P.1987.Sequencefrompicomolequantitiesofproteins elec-troblotted ontopolyvinylidenedifluoride membranes. J. Biol. Chem.

262:10035-10038.

35.Tempst, P., A. J.Link,L. R.Riviere,M.Fleming,and C. Elicone. 1990. Internal sequenceanalysis ofproteinsseparatedonpolyacrylamide gelsatthe