Inhibition of pancreatic phospholipase A2 activity by uteroglobin and antiflammin peptides: Possible mechanism of action

Life Sciences, Vol. 48, pp. 453-464 Printed in the U.S.A.

Pergamon Press

INHIBITION OF PANCREATIC PHOSPHOLIPASE A2 ACTIVITY BY UTEROGLOBIN AND ANTIFLAMMIN PEPTIDES: POSSIBLE MECHANISM OF ACTION. Antonio Facchiano, Eleonora Cordella-Miele, Lucio Miele and Anil B. Mukherjee Section on Developmental Genetics, Human Genetics Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892, Bldg. 10, Rm 9S242 (Received in final form November 27, 1990)

Summary We investigated the possible mechanism of inhibition of porcine pancreatic phospholipase A 2 in vitro by rabbit uteroglobin and by the antiflammin peptides. We optimized the conditions of phospholipase A 2 assay using a deoxycholatephosphatidylcholine mixed micellar substrate and established the activity of these inhibitors under optimized conditions. The results of fluorescence studies and crosslinking experiments indicate that the inhibitors interact with the enzyme in solution and affect the increase in intrinsic fluorescence of phospholipase A 2 observed upon interaction With a mixed micellar substrate. In addition, we identified a sequence similarity between the antiflammin peptides, the putative active region of uteroglobin and a region in pancreatic phospholipase A 2. This region of phospholipase A 2 has been previously identified as being involved in the regulation of dimerization of this enzyme, and is conserved in the pancreatic-type enzymes. Taken together, these observations suggest that uteroglobin and antiflammins interact with porcine pancreatic phospholipase A 2 and this may, at least in part, explain the enzyme inhibitory effect of these molecules observed in vitro. One possible mechanism of this effect may be an interference with the dimerization process of phospholipase A 2 which is associated with interfacial activation. Uteroglobin (UG)¶ is a 15.8 kDa steroid-regulated, homodimeric secretory protein in the rabbit, with immunomodulatory and antiinflammatory properties (1). The antiinflammatory properties of this protein may be, at least in part, due to its phospholipase A 2 inhibitory activity (2). Recently, we reported two nonapeptides derived from UG ,~-helix 3 and lipocortin 1, repeat 3, with potent phospholipase A 2 inhibitory activity in vitro and profound antiinflammatory effect in vivo (3). These peptides correspond to a region of local sequence similarity between UG and lipocortin 1, and are known as antiflammins (AFs). In the present study we investigated the possible mechanism of phospholipase A 2 inhibition by UG and AFs by means of several experimental approaches. Our data suggest that UG and AFs interact with porcine pancreatic phospholipase A2, under our experimental conditions. This phenomenon may explain, at least in part, the previously observed phospholipase A 2 inhibitory effects of these substances in vitro. Materials and methods Materials - rabbit UG was purified as previously described (4). The U G - and lipocortin-lderived oligopeptides AF-1 (MQMKKVLDS) and AF-2 (HDMLKVLDL) respectively, were synthesized and purified by Peptide Technologies, Inc., Washington D.C. Purity was determined by HPLC and amino acid analysis, l-stearoyl-2-[l-14C]arachidonyl-l,3-phosphatidylcholine (sp. act. 58 mCi/mmol) was from Amersham. Arachidonic acid (99% pure, from porcine liver) was from Behring Diagnostic. Porcine pancreatic phospholipase A 2 was purchased from Boehringer (Lot No. 10798523-51) and from Sigma (Lot No. 67F-8197). Silica Gel prechanneled thin layer *To whom all correspondence should be addressed 0024-3205/91 $3.00 +.00

454

Phospholipase A 2 Inhibition by Uteroglobin

Vol. 48, No. 5, 1991

chromatography (TLC) plates, (250 urn), were from Analtech. All other reagents were of analytical grade from either Sigma, Fisher or Fluka. The antiserum against U G was raised in goat; the antibody against porcine pancreatic phospholipase A 2 was developed in rabbit by Assay Research Inc., College Park MD. [126I]UG was prepared by Hazleton Laboratories, Vienna, Va. H P L C - g r a d e water (Fisher) was used throughout this study. Preparation, dilution and storage of e n z y m e and inhibitors - We routinely dialyzed the commercially obtained phospholipase A 2 against 5 changes of 500 volumes of 10 m M Tris HCI, pH 8.0. In the first change, the buffer also contained 0.2 mM phenylmethylsulfonylfluoride (PMSF). After dialysis, the enzyme concentration was calculated by absorbance at 280 nm using an EIc m (1%) of 13.0 (5). The dialyzed enzyme was divided into 50 ~,1 aliquots (final concentration 200 ~,M), flash-frozen in liquid nitrogen and stored at -70 ° C. Each aliquot was slowly thawed on ice prior to its use and the unused portion discarded. Although pancreatic phospholipases A 2 are very stable enzymes and can be stored at 4 ° C as (NH4)2SO 4 precipitates in air tight vials, we noted a slow increase in enzyme activity after opening the sealed vial, even when evaporation was carefully controlled. The storage conditions described above were found to give more reproducible results. Enzyme dilutions for phospholipase A 2 assays were prepared in ice-cold 10 mM Tris HCI, pH 8.0, 1 mM CaCI2, immediately before use. Purified U G was stored as lyophilized aliquots at -70 ° C in the presence of CaSO 4 as desiccant. Aliquots were dissolved in 50% glycerol, 10 m M Tris, pH 8.0, at a final concentration of 1 mM and stored at -20 ° C for about 4 weeks. U G solutions were diluted in ice-cold 10 m M Tris HC1, pH 8.0 and used immediately in the assay. Lyophilized synthetic peptides ( A F - I and A F - 2 ) were stored under nitrogen in sealed ampules at -20 ° C. Storage of these peptides in solution or under non-anhydrous conditions is not recommended since they seem to lose their in vitro phospholipase A 2 inhibitory properties rapidly, once in aqueous environment. We routinely opened one vial of lyophilized peptide immediately before use, and the content was dissolved in ice-cold 10 m M Tris-HCl, pH 8.0, 1 mM CaCI 2. Peptide dilutions were prepared on ice with the same buffer and used immediately. Unused portions of these solutions were discarded. In order to prevent aggregation, concentrated solutions (1 raM) of the peptides were immediately diluted to working concentration. Freezing peptide solutions results in complete loss of activity as observed by us as well as by others (6).

Phospholipase A 2 assay - The assay mixture contained 2 nM phospholipase A2, 1 m M sodium deoxycholate, 10 ~,M [14C]phosphatidylcholine, 100 mM Tris HCI pH 8.0, 100 m M NaCI and 2 mM CaC12 in a total volume of 50 ~,1. For p H - d e p e n d e n c e studies, Tris was replaced by 100 mM morpholinopropanesulfonic acid (MOPS) (pH 6) or 100 mM sodium acetate (pH 4 and 5). [14C]phosphatidylcholine (500 pmoles for each assay tube) was dried under a stream of N 2 for 30-45 minutes and then dissolved in 5 mM sodium deoxycholate (10 ~,1 per assay tube). Fresh deoxycholate solutions were made every day. The phosphatidylcholine-deoxycholate solution was kept at room temperature for 10 minutes and then two volumes of a mix solution containing 250 mM Tris HCI, 250 m M NaCI, 2.5 mM CaCI2, pH 8, was added. The radioactive solution was allowed to stand for at least 1 hour at room temperature before starting the assay. 40 ~,1 of the enzyme solution were mixed with an equal volume of inhibitor solution or buffer, and preincubated for 5 minutes at 37 ° C in an Eppendorf 5320 incubator. Twenty microliter aliquot from this mixture was added to 30 ~,1 of the radioactive substrate to start the reaction. After 30 seconds at 37 ° C the reactions were stopped by adding 50 ~,1 of chloroform:methanol 2:1 (v/v); 50 ~,1 of chloroform and 50 ~,1 of 4M KCI were then added. The samples were mixed by vortex for 30 seconds after each of these three additions. 40 ~,l of the organic phase were spotted onto alternate lanes of prechanneled TLC plates. The plates were preheated for at least 2 hrs at 80 ° C before use. TLCs were run in a chamber saturated with petroleum ether:ethyl ether:glacial acetic acid 70:30:1 for approximately 45 minutes and then developed in a chamber saturated with iodine vapor. Bands comigrating with arachidonate standards were scraped into scintillation vials, and 10 ml of scintillation fluid (Complete Counting Cocktail 3a70B from Research Products International Corp.) were added. Samples were then counted for 5 minutes in an LS-9000 Beckman liquid scintillation counter. Fluorescence studies - Intrinsic fluorescence of phospholipase A 2 in the presence and in the absence of inhibitors was measured with a P e r k i n - E l m e r M P F - 3 0 spectrofluorometer, using a Xenon light source. All the readings were obtained at room temperature. In order to keep the fluorescence proportional to the enzyme concentration, the optical density of all solutions at 280

Vol. 48, No. 5, 1991

Phospholipase A 2 Inhibition by Uteroglobin

455

nm was kept < 0.2. The phospholipase A 2 preparation used for these experiments (Boehringer) appeared homogeneous by sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting (Fig. 1). In addition, this material gave only one UV-absorbing peak when analyzed by Superose 12 gel filtration fast protein liquid chromatography (FPLC) in 25 mM triethanolamine, pH 8.3 and by chromatofocusing on a Mono P-HR20 column between pH 9 and 6 (not shown). Enzyme and inhibitors were dissolved in 10 mM Tris HCI, pH 8.0, 1 mM CaCI 2. The bandpass was 8 nm for both emission and excitation. The procedure used for these experiments was as follows: 15 t,l of 1 mM U G solution or 50 t,1 of 10 mM AFs (or an equal volume of buffer) were added to 1.5 ml of phospholipase A 2 solution, final concentration 3 ~,M, in a fluorometric cuvette. The emission spectrum between 320 nm and 360 nm was recorded with a fixed excitation wavelength of 280 nm. Subsequently, 1 ml of a solution containing 20 mM deoxycholate and 100 ~,M phosphatidylcholine was added to the phospholipase A2-inhibitor mixture and the emission spectrum was recorded as above. Under these conditions, intrinsic fluorescence of porcine pancreatic phospholipase A 2 is primarily due to the only tryptophan present (residue 3). It should be noted that the AFs are devoid of aromatic amino acid residues and UG has no tryptophan. Therefore, we did not expect any significant contribution from UG or AFs. Moreover, we subtracted the contribution of blanks containing inhibitors with or without lipids from each spectrum.

1

5

6

........

A

B

FIG. 1 A: SDS=PAGE of pancreatic phospholipase A 2 (Boehringer). Samples were reduced with 5% t~-mercaptoethanol at 100" C for 10 min in the presence of 1% SDS prior to electrophoresis on a 15% polyacrylamide gel containing 0.1% SDS. Bands were stained with Coomassie blue. h molecular weight standards (BRL, prestained low molecular wight) 2-6: phospholipase A2: 4, 2, 1, 0.5 and 0.25 ~,g respectively. B: Western blot of phospholipase A 2 (0.I t,g) stained with phospholipase A 2 antibody and immunogoldsilver staining. - 25 #1 o f a solution containing 15 ~,g phospholipase A 2 and 15 #g o f UG and trace amounts of [12SI]UG (about 20000 cpm per tube), in 1 mM Tris HCI, pH 8, 10 mM CaCI 2 were preincubated for 1 hour at 37 ° C. Then, 75 ~,1 of 8 mM glutaraldehyde, dissolved in 67 mM triethanolamine, was added. After 1 hour incubation at room temperature, one volume of double strength gel-loading buffer containing 10% ~-mercaptoethanol was added and the samples were boiled for 10 minutes. Ten microliter aliquots were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (7). After electrophoresis, protein bands were transferred to nitrocellulose membranes (8) and the blots were stained with phospholipase A 2 antibody by Crosslinking experiments

456

Phospholipase A 2 Inhibition by Uteroglobin

Vol. 48, No. 5, 1991

means of an immunogold-silver staining kit (Janssen). A f t e r staining, the blots were dried and autoradiographed to detect labeled UG. In another set of experiments, samples were boiled in the absence of p-mercaptoethanol and labeled LTG was detected by autoradiography on dried gels. A Biorad model 620 video densitometer was used to scan autoradioragrams. C o m p a r i s o n o f a m i n o a c i d s e q u e n c e s - The amino acid sequence of U G was compared with that of phospholipases A 2 by means of "GENEPRO" program (Riverside Scientific Enterprises, Seattle, Washington), version 4.1 and manual adjustment. Results

Phospholipase A 2 assay - In order to determine the best conditions to assay putative inhibitors with porcine pancreatic phospholipase A 2 in a mixed micellar system, we evaluated the

i 1 1 1 A

6~-B

i' 3

0

10

20 30 [PL~I, nM

40

Time,mlnuem

1.4 t~ 1.2 "i 1.0

0.7,0 0.5

0.8 o 1~ 0.6

0.4

10.4

0.2

0.2 ae

0.1

1.5

12

0.3

6

pH

8

OI0

10

i

2i 4 lDeoxvchoi~e],mM

e5

E

i

0.9 0.6 0.3

5O 45 4O

0

(Ca++], mM

(4-BPBI,nM

FIG. 2 phospholipase A 2 assay: Panels A, B, and C: concentration- , time- and p H dependence of phospholipase A 2 activity. Panels D and E: dependence of enzyme activity upon the concentrations of deoxycholate and Ca ++. Panel F: inhibitory activity of 4-BPB. Minor modifications to the basic procedure were required to carry out this experiment. A longer preincubation time (10-minutes) was needed to observe the inhibitory activity of 4-BPB (35). 4-BPB was preincubated with PLA 2 in the absence of Ca ++ as Ca ++ protects the enzyme against inactivation (36). The final Ca ++ concentration was kept at 2 nM.

Vol. 48, No. 5, 1991

Phospholipase

A 2 Inhibition by Uteroglobin

457

concentration- and time- dependence of enzyme activity, in addition to the effects of detergent and Ca ++ concentrations and pH. As shown in Fig. 2, under the conditions described above, the rate of phospholipid hydrolysis is linearly dependent upon time, enzyme and Ca ++ concentration (panels A, B, E). The dramatic decrease of enzyme activity above 15 mM Ca ++ is probably due to Ca++-induced aggregation of micelles. No latency period was observed in reaction time courses (Fig. 2B). When the preincubation time of phospholipase A 2 was varied, no significant difference of the activity was noted between 1 and 10 minutes (Fig. 2B, inset). We routinely use 2 nM enzyme, 5 minutes preincubation and 30 seconds reaction and less than 1% of the substrate is hydrolyzed under these conditions. This rate is low enough to prevent significant interference by reaction products. In our assay conditions, we observed a maximal activity between pH 8 and 9 (Fig. 2C). We found that in our conditions the optimum deoxycholate concentration is 1 mM. The increase of activity between 0.5 to 1 mM deoxycholate is probably due to the formation of mixed micelles. In fact, the critical micellar concentration of deoxycholate under similar conditions (pH 7.6 and 0.1 N NaCI) was estimated to be 1 mM (9). The decrease in activity observed above 1 mM deoxycholate may be due to "surface dilution" of the substrate. The inhibitory activity of 4-BPB, a potent, irreversible phospholipase A 2 inhibitor is shown in Fig. 2F. Under the conditions used a maximum inhibitory activity of about 50% was observed above 2 nM. Non-specific proteins (e.g. lysozyme, /~-lactoglobulin, myoglobin and hemoglobin), used as negative controls, did not cause any detectable inhibition of phospholipase A 2 activity between 1 and 100 nM (unpublished data). Inhibition of phospholipase A 2 by U G and AFs under optimized conditions - Fig. 3 shows the dose response curves obtained by preincubating phospholipase A 2 with the indicated inhibitors. At 2 nM concentration of U G and AF-1 30-35% inhibition of phospholipase A 2 activity was

45 40 35

~m 3o O~ -H .-

o

ee-

28 2O 15 10 ~ 1 /

o - o UG

5~ 0~

0

~--i~Antiflammin1 H Antiflammin2 I

I

I

I

I

2

4

6 [Inhibitor], nM

8

10

FIG. 3 Concentration dependence of phospholipase A 2 inhibitory activity of UG, A F - I and AF-2. In all experiments several controls were run, in which enzyme was preincubated with buffer instead of inhibitor. Each point represents the average of several different experiments (n ranging between 13 and 32) each performed in duplicate.

458

A 2 Inhibition

Phosphollpase

5.8

5.4 5

by U t e r o g l o b l n

A

/f~'T~

48, No. 5, 1991

5.8

~

5.4 5

4,8

46

4.2

Vol.

342

4.2

3.8 3.4

B

/~'~\X

3.8 3.4 '

C360

360

iii

Z

LU

U o3 kU

nO .J tl ,,I

> b..J

n-

5.4

5.

5 //

4.6 3.8 3.4

3320

~\

~

~

4.6 3.8 3.4

340 nm

\ I

5.8 5.4 5

4.6 4.2 3.8 3 320

360

3320

//~ f\ 1 345

~

/",

340 nm

360

E = PLA2 + Deoxycholate + Phosphatidylcholine

///~I\\X

- PLA2 Alone

1342 340 nm

360 FIG. 4

Fluorescence spectra of porcine pancreatic phospholipase A 2 in the absence and presence of inhibitors. In all panels, solid lines indicate fluorescence spectra of phospholipase A 2 without mixed micelles; dashed lines represent fluorescence spectra of phospholipase A 2 with mixed micelles. Panel A: phospholipase A 2 alone; panel B: phospholipase A 2 + UG; panel C: phospholipase A 2 + AF-1; panel D = phospholipase A 2 + AF-2; panel E phospholipase A 2 + KVLD. Thick arrows indicate emission maxima and thin arrows indicate the increase of fluorescence at 340 nm observed after the addition of mixed micelles.

Vol. 48, No. 5, 1991

Phospholipase A 2 Inhibition by Uteroglobin

459

observed whereas about 40% inhibition was obtained at 6 nM A F - 2 . When U G was preincubated with the substrate rather than with phospholipase A 2, before adding the mixture to the enzyme, the phospholipase A 2 inhibitory activity of U G was decreased by 70% (unpublished data), as we already observed for A F - I (3). In order to prevent an overestimation of the inhibition due to time dependent loss of activity of diluted enzyme solutions, we routinely ran a reference control every 4-5 experimental points. Additionally, we performed sets of experiments in which the inhibitors were assayed in decreasing order of concentration and reversed this order in other sets of experiments. It is worth mentioning that in these experimental conditions, U G and AFs show an inhibitory effect similar to that of 4-BPB (Fig. 2F). The inhibition was shown to be independent of the concentration of phosphatidylcholine between 1 and 90 ~M (data not shown). studies - The increase of phospholipase A 2 intrinsic fluorescence at 340 nm after the addition of substrate is related to the enzyme activity (10, 11). Thus, we investigated whether U G and AFs had any effect on the fluorescence spectrum of phospholipase A 2 in the presence and in the absence of deoxycholate-phosphatidylcholine mixed micelles. Fig. 4 shows the emission spectra of the enzyme alone and enzyme incubated with inhibitors, before and after the addition of mixed micelles. In the absence of lipids, U G and AFs cause a red shift in the emission maximum of phospholipase A 2. In order to obtain an appreciable signal, the concentrations of enzyme, inhibitor and micelles were increased with respect to the concentrations used in the enzyme assay. In these conditions, the effect of U G was still evident at 1:2 molar ratio of phospholipase A2:UG, whereas the AFs had to be used in 33 fold molar excess to the enzyme. This is probably due to adsorption of AFs to the cuvette walls a n d / o r to the micelles. Importantly, all inhibitors reduced the fluorescence increase at 340 nm, observed after addition of mixed micelles. This increase is reduced 20% by U G , 10% by A F - I , 10% by A F - 2 and it is not affected by the tetrapeptide K V L D . This peptide is common to both A F - I and A F 2, and in itself has no phospholipase A 2 inhibitory activity (3). The reduction in the fluorescence increase at 340 nm is not due to the red shift since identical or higher reductions were measured at emission wavelength maxima. U G and AFs did not have any effect on the intrinsic fluorescence of heat-denatured phospholipase A 2 (not shown).

Fluorescence

Crosslinking experiments - When phospholipase A 2 was reacted with glutaraldehyde and then subjected to S D S - P A G E under reducing conditions and immunoblotting, a number of bands were observed. A similar reaction was performed in the presence of equimolar concentrations of U G , phospholipase A 2 and trace amounts of [19SI]UG (about 800 cpm per lane). In this case an additional band, reacting to a monospecific phospholipase A 2 antibody, appeared at a molecular mass of about 20 kDa (Fig. 5, lane b3). The same band contained radioactive U G (arrow), suggesting that it represents a crosslinked product of one molecule of phospholipase A 2 and one molecule of monomeric U G . Two more additional bands (arrows) of apparent molecular mass of 32 and 43 kDa respectively are visible in the autoradiogram after crosslinking (Fig. 5, lane b4). These bands may represent crosslinked products of: i) one U G dimer and one phospholipase A 2 molecule (about 31 kDa); ii) two U G dimers and one phospholipase A 2 molecule (about 44 kDa) or iii) one U G dimer and two phospholipase A 2 molecules (about 45 kDa). When the reaction was performed in the presence of a 10-fold molar excess of A F - I or A F - 2 over U G , the intensity of the additional bands was clearly reduced (Fig. 5, lanes c6 and d8 respectively). This effect was more pronounced in the case of A F - 1 than A F - 2 . U G residues 1-10, used as a non-specific control, did not affect the intensity of 20 kDa band (Fig. 5, lane el0). A semiquantitative analysis of these data was obtained by densitometric scanning of the autoradiogram. In order to correct for potential artifacts resulting from unequal gel loading, the intensity of the 20 kDa band in each lane was normalized with respect to the sum of the U G monomer and the U G dimer bands in the same lane. Background was measured in lane f12, where no 20 kDa band was observed, and subtracted from all data. The corrected relative intensities of the 20 kDa bands in lanes b4, c6, d8 and el0 are 100, 55, 78 and 92 respectively. AFs did not affect the intensity of other bands in the Western blot, except the 20 kDa band, suggesting that the region of U G corresponding to A F - 1 may be involved in the crosslinking. Furthermore, A F - 2 , which is similar but not identical to A F - 1 , does not affect the 20 kDa band as effectively as A F - I . These observations suggest that U G Lys 43 and its counterpart in A F - I and A F - 2 may be involved in crosslinking. Fig. 6 shows the results of reacting glutaraldehyde with U G and trace amounts of [12SI]UG in the absence (lane a) or in the presence of phospholipase A 2 (lane b). The reaction products were analyzed by S D S - P A G E under non-reducing conditions. In lane a the band of

460

A 2 Inhibition by Uteroglob±n

Phosphol±pase

Vol. 48, No. 5, 1991

apparent molecular mass o f 32 kDa represents a crosslinked complex o f two U G dimers; the 16 kDa band is the U G dimer. In lane b two additional bands (arrows) are visible w i t h molecular masses of 31 kDa and 44 kDa respectively. The 31 kDa band may correspond to a crosslinked complex o f one U G dimer and one molecule o f phospholipase A 2. The 44 kDa band may represent a complex o f one molecule o f phospholipase A 2 w i t h two U G dimers, or a complex of one U G dimer w i t h two phospholipase A 2 molecules.

a 1

b 2

3

c 4

5

d 6

7

e 8

a

f 10

11

12

/k3kDa -25kDa

--

18kDa--

FIG. 5 SDS-PAGE of phospholipase A 2 - U G crosslinked complexes under reducing conditions. Samples were reduced with 5% /~-mercaptoethanol at 100* C for 10 min in the presence of 1% SDS prior to electrophoresis on a 15% polyacrylamide gel containing 0.1% SDS. Bands were transferred to a nitrocellulose membrane. This membrane was first stained with phospholipase A 2 antibody and then autoradiographed, a: phospholipase A 2 alone; b: phospholipase A2+ UG; c: phospholipase A2+ UG + AFI; d: phospholipase A2+ U G + AF2; e: phospholipase A 2 + UG + non specific peptide (UG residues 1-10); and f = U G alone. Odd numbers indicate the immunostaining and even numbers indicate the autoradiogram of each lane.

a kDa--

b i

.g-.

kDa--

FIG. 6 SDS-PAGE of phospholipase A 2 - U G crosslinked complexes under non-reducing conditions. Samples were treated at 100° C for 10 rain in the presence of 1% SDS but without reducing agents prior to electrophoresis on a 15% polyacrylamide gel. The gel was dried and autoradiographed, a: U G alone; b: UG + phospholipase A 2. The arrows indicate the additional protein bands observed when [lzSI]UG was crosslinked with phospholipase A 2.

Vol. 48, No. 5, 1991

Phospholipase A 2 Inhibition by Uteroglobin

461

Comparison of amino acid sequences between U G and phospholipnse A 2 - Previously, we were

unable to identify any sequence similarity between UG and phospholipase A 2 using FASTP (3). However, by using GENEPRO 4.1 program and by manually introducing gaps, we have now identified a 35% similarity (24 identities out of 70) between the mature UG sequence and residues 8-78 of porcine pancreatic phospholipase A 2 (Fig. 7a). This is consistent with the striking similarity in the hydropathy profiles of the two proteins we already reported (3). Moreover, on the basis of the atomic coordinates, the calculated molecular surfaces of UG and phospholipase A 2 from Crotalus a t r o x have been suggested to be almost superimposable (12).

A i PigPLA2

io

M K

AI

io

8

2o

~.MDF

NYGCYCG

~

~

GS

3o

VD - -

RCC

40 --

40

DNCYR SO

~o

....

KFLV6O

6o Y--

70 Y--

m

B Rabbit UG Human PLA2 Bovine PLA2 Equine PLA2 Porcine PLA2

40-47 54-60 54-60 54-60 54-60

QMKKVLDS QA KKLDS QA K K L DS QA K ELSS DA KNLDS

FIG. 7 Amino acid sequence comparison between UG and phospholipase A 2. a: alignment of mature UG with porcine pancreatic phospholipase A 2. Gaps were manually introduced to optimize the alignment. Dots indicate conservative substitutions; b: local similarity between UG residues 40-47 and phospholipases A 2 from human lung, bovine pancreas, equine pancreas and porcine pancreas, residues 54-60. More interestingly, when the region of UG corresponding to AF-I was matched to the sequence of pancreatic-type phospholipases A 2, a high local similarity was identified with a specific region of these enzymes (Fig. 7b). This region of phospholipase A 2 contains Lys 56, which has been shown to be autocatalytically acylated, a prerequisite for dimerization and consequent interfacial activation of pancreatic phospholipase A 2 (13). Additionally Ala in position 55 of phospholipase A 2, an almost invariant residue, allows few substitutions. One of these is Met, matching in this case with Met 41 of UG. Discussion We have optimized the experimental conditions to study porcine pancreatic phospholipase A 2 inhibition by UG and AFs in a mixed micellar assay. These conditions may also be useful to evaluate other putative polypeptide inhibitors of phospholipase A 2. The use of the porcine pancreatic enzyme as a first approximation model to study phospholipase A 2 inhibitors may be justified by the following considerations: first, among the commercially available phospholipases A2, porcine pancreatic phospholipase A 2 is evolutionarily closer to physiologically relevant mammalian enzymes than snake or bee venom enzymes; second, in cellular signal transduction via the G-protein cycle, it has been shown that exogenously added porcine pancreatic phospholipase A 2 can substitute for the G-protein regulated intracellular enzyme involved in this process, in RBL cells (14); third, gene cloning studies from several tissues suggest that pancreatic type phospholipase A 2 is not restricted to pancreas only. For instance, the human lung phospholipase A 2 is identical to the pancreatic enzyme (15) and human, rat and pig non-pancreatic phospholipases A 2 have been found to have very similar sequences to the pancreatic type enzyme (16). Although in most cases the non-charged detergent Triton X-100 has been used, we preferred deoxycholate/phosphatidylcholine micelles, because this mixture more closely resembles

462

Phospholipase A 2 Inhibition by Uteroglobin

Vol. 48, No. 5, 1991

the physiological substrate of pancreatic phospholipase A2. Moreover, the porcine pancreatic enzyme has a very low affinity for mixed micelles containing Triton (17) as well as for phospholipid bilayers and vesicles, and it has a clear preference for negatively charged substrates. In our hands, substituting Triton X-100 for deoxycholate, under otherwise identical conditions, caused a three orders of magnitude decrease in phospholipase A 2 activity (unpublished data). With most substrates, except for negatively charged mixed micelles, the binding and "penetration" of this phospholipase A 2 into the interface is relatively slow (18, 19). Under such conditions it is easier to observe artifactual inhibition by substances that simply bind to phospholipids or adsorb to lipid/water interfaces. On the other hand, conditions resulting in instantaneous interfacial activation of phospholipase A2, in the presence of a large molar excess of detergent, should minimize non specific effects due to lipid binding. This is particularly true if the putative inhibitors are tested at very low molar concentrations, in the same range of the enzyme concentration. In order to obtain reproducible results in the phospholipase A 2 assay, we found several conditions to be essential. These include the use of HPLC-grade water and ultrapure reagents for all solutions, the filtration of the buffers through membranes with pore size 0.22 ~m, the storage conditions of dialyzed phospholipase A 2, the use of polypropylene tubes and ice-cold buffer for phospholipase A 2 dilution. Moreover, the time intervals between the thawing of phospholipase A 2 stock solution in ice, the preparation of enzyme dilutions and their use in the assay must be as consistent as possible, within three hours. Under these conditions, we established more firmly the phospholipase A 2 inhibitory properties of UG and AFs in nanomolar concentrations, although the average percent inhibition obtained was slightly lower than previously reported (3). Failure to adhere to the conditions described for preparation and storage of the inhibitors and for phospholipase A 2 assay conditions may result in the lack of reproducibility or inability to detect inhibitory activity (20). Particularly the AFs have been shown to be unstable in vitro (6, 21, 22). At least two mechanisms seem to contribute to this instability: oxidation of Met residue(s) (6) and an as yet unclarified physical phenomenon that leads to complete inactivation of the peptides upon freezing in solution (6). The latter phenomenon may partly be due to peptide aggregation or to adsorption to the walls of the tube. Our data suggest that: i) UG and AFs interact with phospholipase A 2 in solution and may directly or indirectly increase the polarity in the microenvironment of the side chain of Trp 3; ii) in the presence of UG or AFs, the increase in phospholipase A 2 fluorescence observed upon addition of substrate is reduced; iii) UG can be crosslinked to phospholipase A 2 in the absence of lipids, and this reaction is inhibited in the presence of an excess of AFs; iv) AFs, and the corresponding putative active site of UG show a sequence similarity to a conserved region in pancreatic-type phospholipases A 2 which contains a Lys involved in the control of enzyme dimerization. Taken together, these data suggest that at least one possible mechanism through which these inhibitors affect phospholipase A 2 activity in vitro may be by interacting with the enzyme, and possibly interfering with the self-association of phospholipase A 2 which accompanies interracial activation. Other preliminary observations and theoretical considerations support this hypothesis: i) [125I]UG does not bind E. coli membranes in the presence or absence of Ca ++ (not shown); ii) when analyzed by steady-state kinetics, the inhibition seems to be noncompetitive with respect to phosphatidylcholine between 1 and 90 ~,M; iii) other protein effectors that bind and inhibit hydrolases have been shown to be structurally similar to the enzymes they inhibit. Mancheva et al. (23) have reported sequence similarity between bulgarian viper venom phospholipase A 2 and its physiological inhibitor vipoxin. A local amino acid sequence similarity of 35% was also found between porcine -,-amylase and its inhibitor Z-2685 from Streptomyces parvulus FH-1641 (24). Although our data do not support any evolutionary relationship between UG and phospholipases A 2, the observed common structural features may suggest a parallel with bulgarian viper phospholipase A 2 and its inhibitor (23), and a-amylase and its inhibitor (24). The putative mechanism of action of U G and AFs in our experimental system shows important differences from the suggested mechanism of the in vitro inhibition of phospholipase A 2 by lipocortin 1 and other related proteins (25, 26). Although these proteins do display a weak interaction with pancreatic phospholipase A 2 in solution (27, 28), they also interact with various phospholipids, and the latter interaction is necessary for their in vitro antiphospholipase activity (25, 26). It should be mentioned that, despite its structural similarity with lipocortin 1 repeat 3, UG does not possess the "endonexin consensus~ (29) and does not seem to bind phospholipids in vitro. Thus, UG cannot be considered as a member of the "lipocortin family" in evolutionary terms. It has been recently shown that AFs inhibit human neutrophil phospholipase A 2 activity and the generation of PAF, as well as neutrophil

Vol. 48, No. 5, 1991

Phospholipase A2 Inh±bitlon by Uteroglobin

463

chemotaxis, aggregation and intradermal Arthus reaction (6). AFs have been also demonstrated to inhibit tumor necrosis factor-induced and phagocytosis-induced araehidonate release from polymorphonuclear leukocytes in vitro (22). In addition, AFs have been shown to have a potent antiinflammatory effect in endotoxin-induced uveitis in rats (30). Vostal et al. (31) have shown that A F - I inhibits platelet aggregation and serotonin secretion induced by ADP. UG had previously been shown to inhibit neutrophil and monocyte ehemotaxis and platelet aggregation in vitro (32, 33). Like the antiinflammatory effect (3), the effect of AFS on platelet aggregation is drastically reduced in the presence of 10-4 M arachidonic acid, and is presumably due to phospholipase A 2 inhibition (31). The anti-inflammatory effect of both A F - I and AF-2 in the carrageenan-induced footpad edema has also been independently confirmed by Ialenti et al. (34) and shown to be most pronounced in the late phase rather than the early phase of the inflammatory reaction. On the basis of this observation and due to the ineffectiveness of these substances on dextran-induced edema, which does not involve phospholipase A2 activation, these authors conclude that "the antiflammins inhibit the generation of eicosanoids probably by the inhibition of phospholipase A2." (34). However, it remains to be clarified whether the mechanism of the observed inhibitory effect on porcine pancreatic phospholipase A 2 in vitro and the mechanism(s) through which UG and AFs prevent release of arachidonate and exert their various biological effects are identical. Acknowledaements We thank Drs. Sop I1 Chung and Hao-Chia Chen for critical review of the manuscript and valuable suggestions. We are grateful to the Cystic Fibrosis foundation for partial support of this project. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

L. MIELE, E. CORDELLA-MIELE, and A. B. MUKHERJEE, Endocr. Rev. [ 474-490 (1987). S.W. LEVIN, J. D. BUTLER, P. WIGHTMAN, U. K. SCHUMAKER and A. B. MUKHERJEE, Life Sci. 3_881813-1819 (1986). L. MIELE, E. CORDELLA-MIELE, A. FACCHIANO and A. B. MUKHERJEE, Nature 335 726-730 (1988). A.B. MUKHERJEE, R. E. ULANE and A. K. AGRAWAL, Am. J. Reprod. Immunol. 2135-141 (1982). M . C . F . VAN DAM-MIERAS, A. J. SLOTBOOM, W. A. PIETERSON and G. H. DE HAAS, Biochemistry 1_~45387-5393 (1975). G. CAMUSSI, C. TETTA, F. BUSSOLINO, and C. BAGLIONI, J. Exp. Med. 17___L1913-937(1990). U . K . LAEMMLI, Nature 22____7_7680-685(1970). H. TOWBIN, T. STAEHELIN, and J. GORDON, Proc. Natl. Acad. Sci. USA 764350-4354 (1979). D . M . SMALL, The Bile Acids: Chemistry. Phvsioloav and Metabolism. P. P. Nair and D. Kritchewsky (eds), vol 1, 249-356, Plenum Press, New York (1971). J. D. BELL and R. L. BILTONEN, J. Biol. Chem. 264225-230 (1989). J. D. BELL and R. L. BILTONEN, J. Biol. Chem. 2~i412194-12200 (1989). I. MORIZE, E. SURCOUF, M. C. VANEY, Y. EPELBOIN, M. BUEHNER, F. FRIDLANSKY, E. MILGROM and J. P. MORNON, J. Mol. Biol. 194725-739 (1987). A.G. TOMASSELLI, J. HUI, J. FISHER, H. ZURCHER-NEELY, I. M. REARDON, E. ORIAKU, F. J. KEZDY, and R. L. HEINRIKSON, J. Biol. Chem. 26410041-10047 (1989). V. NARASIMHAN, D. HOLOWKA, B. BAIRD, J. Biol. Chem. 2641459-1464 (1990). J. J. SEILHAMER, T. L. RANDALL, M. YAMANAKA, and L. K. JOHNSON, DNA 5519-527 (1986). J. J. SEILHAMER, T. L. RANDALL, L. K. JOHNSON, I. HEINZMANN, R. S. KLISAK and A. J. LUSIS, J. Cell. Biochem. 39327-337 (1989). A. J. SLOTBOOM, R. VERGER, H. M. VERHEIJ, P. H. M. BAARTSMAN, L. L. M. VAN DEENEN, andG. HDE HAAS, Chem. Phys. Lipids 17128-147 (1976). R. VERGER, Ann. Rev. Biophys. Bioeng. 5477-177 (1976). A. J. SLOTBOOM, H. M. VERHEIJ, and G. H. DE HAAS, Phosoholinids. New Comprehensive Biochemistry, J. N. Hawtorne and G. B. Ansell (eds), vol. 4, 359-434, Elsevier North Holland, Amsterdam (1982).

464

Phospholipase A2 Inhibition by Uteroglobin

Vol. 48, No. 5, 1991

20. J. VAN BINSBERGEN, A. J. SLOTBOOM, A. J. AARSMAN, and G. H. DE HAAS, FEBS Left. 24___!293-297 (1989). 21. A. B. MUKHERJEE, E. CORDELLA-MIELE, A. FACCHIANO, L. MURTY, M. J. KIM, L. MIELE, Biochemical Asoects on the ImmunoDathol~v of Reoroduction. G. Spera, A. B. Mukherjee, G. Ravagnan, S. Metafora (eds), 65-81, Acta Medica, Rome (1989). 22. G. CAMUSSI, C. TETTA, E. TURELLO and C. BAGLIONI, CytQkines and Lioocortins in Inflammation and DifferenIi~ttign, M. Melli and L. Parente (eds), 69-80, Wiley-Liss, New York, (1990). 23. I. MANCHEVA, T. KLEINSCHMIDT, B. ALEKSIEV, and G. BRAUNITZER, Hoppe Seyler's Z. Physiol. Chem. 365 885-894 0984). 24. O. HOFMANN, L. VERTESY, G. BRAUNITZER, Biol. Chem. Hoppe-Seyler 366 1161-1168 (1985). 25; F. F. DAVIDSON, E. A. DENNIS, M. POWELL and J. R. GLENNEY, J. Biol, Chem. 262 1698-1705 (1987). 26. F. F. DAVIDSON, M. D. LISTER, and E. A. DENNIS, J. Biol. Chem, 2655602-5609 (1990). 27. N.G. AHN, D. C. TELLER, M. J. BIENKOWSKI, B. A. McMULLEN, E. W. LIPKIN and C. DE HAEN, J. Biol. Chem. 26318657-18663 (1988). 28. L. PARENTE, and R. J. FLOWER Life Sci. 361225-1231 (1985). 29. R. H. KRETSINGER and C. E. CREUTZ, Nature 320573 (1986). 30. CHAN, C.C., NI, M., MIELE, L., CORDELLA-MIELE, E., FERRICK, M., MUKHERJEE, A. B., NUSSENBLATT, R. B., Arch. Ophthalmol., in press. 31. J. VOSTAL, A. B. MUKHERJEE, L. MIELE, L. and N. R. SHULMAN, Biochem. Biophys. Res. Commun. 16527-36 (1989). 32. G. VASANTHAKUMAR, R. MANJUNATH, A. B. MUKHERJEE, H. WARABI and E. SCHIFFMAN, Biochem. Pharmacol. 37389-394 (1987). 33. R. MANJUNATH, S. W. LEVIN, K. K. KUMAROO, J. DEB. BUTLER, J. DONLON, H. McDONALD, R. FUJITA, U. K. SCHUMAKER and A. B. MUKHERJEE, Biochem Pharma¢ol. 36 741-746 (1987). 34. A. IALENTI, P. M. DOYLE, G. N. HARDY, D. S. E. SIMPKIN, and M. DI ROSA, Agents and Actions 2..9948-49 (1990). 35. J.J. VOLWERK, W. A. PIETERSON and G. H. DE HAAS, Biochemistry 131446-1454 (1974). 36. W. A. PIETERSON, J. J. VOLWERK and G. H. DE HAAS, Biochemistry, 13, 1439-1445 (1974).