ELSEVIER FEMS Microbiology Letters I35 (I 996) 143- 147

Effect of Vibrio choZerue non-01 protease on lysozyme, lactoferrin and secretory immunoglobulin A

Claudia Toma a> * , Yasuko Honma a, Masaaki Iwanaga a-b

’ Deparrment of Bacteriology, Uniwrsir? of the Ryukyus. Nishihuro, Okina,co YO3-01, Joptrn

h Resetrrch lnsrifure qf Conrprehensil~e Medicine. Faculr\ of Medicine, Uniwr.si?y ofthe Ryky~s. Nishihura, Okinmw 903-01. Jopat!

Received 23 October 1995; accepted 10 November 1995

Abstract

The effect of Vibrio cholerue non-01 protease on host defense proteins (lysozyme, secretory immunoglobul~in A and lactoferrin) was studied in relation to its virulence mechanism. The proteins treated with the protease were analysed by SDS-PAGE. There was no influence of the protease on lysozyme. The protease cleaved lactoferrin into two fragmgnts of SO kDa and 34 kDa. N-terminal amino acid sequencing of these fragments revealed that the cleavage site was near the hinge region, between serine 420 and serine 421. This cleavage could affect the transition from open to closed configuration which is involved in iron binding and release. The anti-bacterial activity of lactoferrin was not affected by protease tireatment. Secretory immunoglobulin A yielded a 42-kDa protein as the cleavage product. The susceptibility of secretory immuno-

globulin A to V. cholerae non-01 protease suggests a mechanism by which bacteria might evade the effect of this immunoglobulin.

Keword.~: Vibrio cholerae non-01; Protease; Lysozyme; Lactoferrin; Secretory immunoglobulin A

1. Introduction

Vibrio cholerue non-01 protease was first puri-

fied and reported by Honda et al. 111, and its compar- ison with the V. cholerue 01 soluble hemag- glutinin/protease revealed that they were immuno- logically and physicochemically identical. Many in-

vestigators have postulated the possible roles of V. cholerue protease. At first, it was thought to be involved in bacterial colonization, because digestion of mucin and/or fibronectin might facilitate vibrio approximation to the eukaryotic cell surface [2].

_ Corresponding author. Tel.: + 81 (98) 895 3331; Fax: + 81

(98) 895 295 I ; E-mail: k950417@med.u-ryukyu.ac.jp.

Nevertheless, recent experiments using protaase-defi- cient mutants showed that they still had the ability to attach to cultured human intestinal epithelial cells [3]

and the protease was likely to play a role io detach- ment more than in attachment [3]. V. cholarm pro- tease was also considered to nick and activhte the A

subunit of cholera toxin (CT) [2]. However, ;the wide distribution of the protease among non-CT-produc- ing V. cholerue non-01 strains suggested more bio-

logical roles. Furthermore, proteins that may partici- pate in host defense against cholera, such as mucin, fibronectin and lactoferrin, were digested ob cleaved

by V. cholerue 01 protease [2]. Mucosal secretions have an anti-bacterial arma-

mentarium very different from that of serum. There

0378.1097/96/$12.00 0 1996 Federation of European Microbiological Societies. All rights reserved SSDI 0378- 1097(95)00463-7

are high concentrations of secretory immunoglobulin

A (sIgA), lysozyme, and lactoferrin in the secretions. The killing of V. cholerae by lactoferrin [4], and a bactericidal synergistic effect between lactoferrin and

lysozyme [5] or lactoferrin and sIgA [6] have been reported. During in vivo infection, the protease re- leased from bacteria could induce an irreversible degradation or inactivation of locally produced pro-

teins which are potentially important in host defense. Pseudomonas aeruginosa elastase digested lysozyme

[7], cleaved lactoferrin [S] and degraded IgA [9]. Although not clear, it is postulated that P. aerugi-

nosa elastase could be an important virulence factor. V. cholerae protease and P. aeruginosa elastase

were immunologically and functionally related [81. Therefore, V. cholerae non-01 protease could also

be a factor that decreases the host mucosal defense and thus might enhance the pathogenic potential.

In the present study, we investigated the effect of non-01 protease on host defense proteins such as

lactoferrin, lysozyme, and sIgA.

2. Materials and methods

2. I. Bacterial strain

Vibrio cholerae non-0 1, strain 93Ag 13, isolated from a cholera-like patient in 1992 in Argentina

(provided by Dr. M. Rivas, Instituto National de Microbiologia “C. Malbran”), was used for protease purification. This strain did not produce cholera-like

toxin.

2.2. Protease pur@cation

Bacteria were precultured in trypticase soy broth (TSB) (Difco, Detroit, MI) with shaking at 37°C for

4 h. The precultured bacteria were inoculated into 500 ml of TSB in a 3-1 Erlenmeyer flask, and incubated for 20 h with shaking at 30°C. The cell-free culture supernatant was fractionated with ammonium sulfate. The 40-55% ammonium sulfate-insoluble material was suspended in 0.02 M Tris . HCl buffer (pH 7.0) and dialysed against the same buffer. The dialysed material was applied to a Biogel A5m col- umn (Bio-Rad Laboratories, Richmond, CA) and the fractions showing protease activity were applied to a

TSK gel G-3000 SW column (Tosoh, Tokyo, Japan)

on HPLC as described by Ichinose et al. [IO].

2.3. Protease acticit?;

Human lactoferrin (LF) (Sigma Chemical Co.. St. Louis, MO), human lysozyme (LYJ (Sigma) and

human sIgA (Organon Teknika N.V., Cape1 Prod- ucts, NC) were reacted with purified protease at 37°C for an appropriate time (1 - 16 h) at an enzyme- to-substrate molar ratio of I:800 for LF, I :800 and

I:5 for LY, and I:300 and I :4 for sIgA. The reaction mixtures were analysed by Laemmli’s sodium-

dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) [ 111 with a gel concentration of 9% for LF and sIgA, and 15% for LY. The proteins were

visualized by Coomassie brilliant blue staining.

2.4. Amino acid sequence analysis

The peptides of LF that appeared on SDS-PAGE were blotted onto a polyvinylidene difluoride (PVDF) membrane (Applied Biosystems, CA). The N-termi-

nal amino acid sequence of the blotted proteins were analysed by automated Edman degradation on a Shi-

madzu PSQ-I protein sequencer. The cleaving site was determined by comparing with the whole se-

quence reported by Anderson et al. [ 121.

2.5. Effect of protease on lactoferrin actki8

An inoculum of 5 X 10’ cfu of bacteria was added to 200 ~1 of Bactopeptone 1% containing 330

PM Zincov (a protease inhibitor; Calbiochem, La Jolla, CA). The medium was supplemented with 6

mg ml-’ of LF, or with LF treated with protease.

The mixtures were incubated at 37°C aliquots re- moved at different times and cfu determined.

3. Results and discussion

SDS-PAGE of LY treated with protease revealed that lysozyme was not affected by the protease (data not shown). Although incubation of P. aeruginosa

elastase with human lysozyme resulted in disappear-

Ml 2 34 56

175 K

83

62

47.5

32.5

Fig, I. Coomassie brilliant blue-stained SDS-PAGE 97~ of LF and

sIgA. Lane M, molecular mass markers; lanes I and 2. human LF:

lanes 3-S. slgA. Lane 6. purified protease. Lanes 1 and 3. treated

with control buffer; lanes 2. 3 and 5, treated with protease: lane 2.

I :8OO: lane 4. 1:300: lane 5. 1:3. All reaction mixtures were

incubated at 37°C: I h for lanes I and 2: 3 h for lanes 3 and 3; and

overnight for lane S. Steined hands in lane 3 (from the top): sIgA.

IgA. SC. H chain. L chain.

ante of the native lysozyme band on SDS-PAGE [7],

V. cholerue non-01 protease had no effect in this system.

3.2. Luctoferrin

Treatment of human LF with protease (I $00 for I h at 37°C) yielded two fragments of SO kDa and 34 kDa (Fig. 1. lane 2). Finkelstein et al. 121 also

detected two fragments of 55 kDa and 37 kDa when

LF was treated with V. cholrme 01 protease. N- terminal amino acid sequencing of the two fragments

showed that the cleaving site was in the C-lobe between serine 420 and serine 421 (Fig. 2).

LF is a glycoprotein composed of two lobes

(N-lobe and C-lobe). Each of them has an iron-bind- ing site in a deep cleft between two domains. The

C-lobe is more stable and releases iron slower than the N-lobe [ 121. In the C-lobe. amino acids 407-443 form a ‘backbone‘ strand running the whole length of the lobe that has one interruption in the form of an

1 10 20 N-lobe GRRRSVQWCAVSNPEATKCFQWQRNMRKVRGPPVSCLKRDSPIQCIQAIAENR

C-lobe ARRAR""WCA"GEQELRKCNQ~SGLS~----GSVTCSSASTTEDCIALVLKG~

310 350 360 370 380

60 70 80 90

N-lobe ADAVTLDGGFIYEAGLAPYKLRPVAAEVYGTE----------RQPRTHYYAVA

C-lobe A D A M S L D G 1; Y V Y T A G K C - - G L " P V L A E N Y K S Q Q S S__D.P_D_P_N_~._.V____~ P V E G I L A " A

390 400 4 1 0 420 430

100 110 120 130 140

N-lobe VVKKGG-SFQLNELQGLKSCHTGLRRTAGWNVPIGTLRPFLNWTGPPEPIEAA

C-lobe VVRRSDTSLTWNSVKGKKSCHTAVDRTAGWNIPMGLLFNQTGSC-------K

440 450 460 470 480

150 160 170 180 190

N-lobe "ARFFSASCVPGADKGQFPNLCRLZAGT--GENKCAFSSQEPYFS~SGAFKCL

C-lobe FDEYFSQSCAPGSDPR--SNLCAL_IGDEQGENKCVPNSNERYYG~TGAFRCL

490 500 510 520 530

200 210 220 230 240 N-lobe KDGAGDVAFIRESTVFEDLSD--------EAERDEYELLCPDNTRKPVDKFKD

c-lobe AENAGDVAFVKDVTVLQNTDGNNNEA~AKDLKLADFALLCLDGKRKPVTEARS 540 550 550 570 580

250 260 270 280 290

N-lobe CHLARVPS~AVVARSVNGKEDAIWNLLRQAQ~KFGKD---KSPKFQLFGSPSG

C-lobe CHLANAPNRAVVSRM--DKVERLKQVLLHQQAKFGRNGSDCPDKFCLFQS~--

590 600 610 620 630

300 310 320 330 140

N-lobe QKDLLFKDSAIGFSR"PPRIDSGLYLGSGYFTAIQ~LRKSEKE"AARRAR C-lobe TKNLLFNDNTECLARLHGKTTYEKYLGPQYVAGITNLKKCSTSPLLEACEFLRK

640 650 660 670 680 690

Fig. 3. Amino acid sequence of human lactoferrin. Alignment of the N-terminal and C-terminal halves is based on the Uperposition of their

three-dimensional structures as reported by Anderson et al. [I?]. The fragments obtained after protease treatement (I:800 for I h) were transferred to a PVDF membrane and the N-terminal amino acid sequence was determined. Single underline. N-terminal sequence of the

SO-kDa fragment. Dashed underline, N-terminal sequence of the 34.kDa fragment. Residues involved in iron-binding are indicated by

xhaded Ictters.

146 C. Toma et al./ FEMS Microbiology Letters 135 (19%) 143-147

inserted loop (amino acids 416-426). This arrange- ment might give some flexibility in opening and

closing the inter-domain cleft [ 121. Iron binding oc- curs to the open form to give an intermediate in which the ligand is only bound to one domain. After

that, closure of the domains occurs to allow the ligand to bind to both domains, giving the closed,

ligated structure. Therefore, cleavage at this position

could affect the transition from open to closed con- figuration that is involved in iron binding and release [ 121. Although the two lobes present a similar

polypeptide chain conformation, the region that was cleaved by protease has no superposition in the N-lobe (Fig. 2). This could explain why the N-lobe

was not cleaved by the protease in a homology

region. The anti-bacterial domain of LF was identified

near the N-terminus (amino acids 18-47) and it was

demonstrated that anti-bacterial activity of lactoferrin is independent of its iron-binding property [ 131.

Cleavage at the N-terminus might not induce any change in the molecular mass of the whole LF and

therefore might not be detected in the SDS-PAGE. To investigate if the anti-bacterial domain was inac- tivated, the growth of bacteria in media containing

LF or LF treated with protease were compared. The anti-bacterial activity of LF was not affected by protease treatment (Fig. 3). Although non-cleaved LF was left after the protease treatment. the effect of

LF on bacterial growth is dose-dependent and some change in the growth curve would be expected if the

anti-bacterial domain was affected.

0 1 2 incub%on t&e (h) 5

6 7

Fig. 3. Effects of lactoferrin (LF) 6 mg ml-’ and lactoferrin 6 mg

ml-’ treated with protease (LF+ P) on the growth of V. cholerae

non-01 strain in bactopeptone 1% containing Zincov 330 PM.

3.3. Secretor?, immunoglobulin A

There was no change in the secretory component

(SC) and heavy and light chains of the sIgA after treatment with protease with an enzyme-to-substrate ratio of 1:300. Nevertheless, sIgA disappeared and a

42-kDa protein was detected (Fig. 1, lane 4). When protease concentration and incubation time were in- creased (1:4, overnight) both the 42-kDa protein and

SC disappeared (Fig. 1, lane 5). Finkelstein et al. [2] reported a decrease in trichloroacetic acid-precipi- table radioactive material when sIgA was treated

with V. cholerae 01 protease. However, they did not observe these changes in the SDS-PAGE. Our results suggest that protease degraded SC, yielding a 42-kDa

fragment at low enzyme concentration. When the enzyme concentration and incubation time were in- creased, further proteolysis of the SC and 42-kDa

fragment occurred. Resistance to proteolytic degradation of sIgA due

to the binding of SC has been reported [14]. The resistance of sIgA to intestinal proteases makes anti- bodies of this isotype uniquely well suited to protect intestinal mucosal surfaces. Therefore, the suscepti-

bility of sIgA to V. cholerae non-01 protease re- ported here might represent a factor that allows the bacteria to escape from the host immune response

and to establish infection.

Acknowledgements

We thank Tokushu Meneki (Tokyo, Japan) for generously providing bovine lactoferrin used in the preliminary experiments.

References

[I] Honda, T.. Lertpocasombat, K., Hata, A., Miwatani, T. and

Finkelstein, R.A. (I 989) Purification and characterization of

a protease produced by Vibrio cholerae non-0 I and compar-

ison with a protease of V. cholerue 01. Infect. Immun. 57.

2799-2803.

[2] Finkelstein, R.A., Boesman-Finkelstein, M. and Holt, P.

(1983) Vibrio cholerue hemagglutinin/lectin/protease hy-

drolyzes fibronectin and ovomucin: F.M. Burnet revisited.

Proc. Natl. Acad. Sci. USA 80, 1092-1095.

[3] Finkelatein. R.A., Boesman-Finkelstein, M., Chang, Y. and

HPse. C. (1992) Vihrio cholrrtrr hemmaglutinin/protease,

colonial variation. virulence, and detachment. Infect. Immun.

60.472-478.

[4] Arnold. R.R.. Cole, M.F. and McGhee, J.R. (1977) A bacte-

ricidal effect for human lactoferrin. Science 197, 263-265.

[5] Elliaon III, R.T. and Giehl. T.J. (1991) Killing of Gram-nega-

tive bacteria hy lactoferrin and lysozyme. J. Clin. Invest. XX.

1080-1091.

[6] Boesman-Finkelstein, M. and Finkelstein, R.A. (1985) An-

timicrobial effects of human milk: inhibitory activity on

enteric pathogens. FEMS Microbial. Lett. 27, l67- 174.

[7] Jacquot. J.. Tournier. J.M. and Puchelle. E. (1985) In vitro

evidence that human airway lysozyme is cleaved and inacti-

vated by Psrudorw~ms urru~inosa elastase and not by hu-

man leukocyte elastase. Infect. Immun. 47. 555-560.

[8] H&e. C. and Finkelstein. R.A. (1990) Comparison of the

Vihno chlrrtru hemagglutinin/protease and the Psew

do1nonn.r cwru~irrosrr elastase. Infect. Immun. 58, 401 I-4015.

[9] Heck. L.W., Alarcon. P.G.. Kulhavy. R.M., Morihara, K..

Russell, M.W. and Mestecky. J.F. (1990) Degradation of IgA

proteins by Pseudornonns neruginoso elastase. J. Immunol.

144, 2253-2257.

[IO] Ichinose, Y.. Ehara. M. and Utsunomiya A. (1992) Purifica-

tion of protease from V. dderrrr 01 and its partial charac-

terization. Trop. Med. 34. I2 I - 125.

[I I] Laemmli, U.K. (1970) Cleavage of structural protein5 during

the assembly of the head of bacteriophage T4. Nature 227.

680-685.

[l2] Anderson. B.F.. Baker, H.M., Norris. GE. Rice. D.W. and

Baker. E.N. (I 989) Structure of human lactoferrin: cry$tallo-

graphic structure analysis and refinement at 2.X A resolution.

J. Mol. Biol. 209. 71 l-734.

[I31 Bellamy. W.. Takase, M.. Yamauchi. K.. Wakabayashi, H.,

Kawahe. K. and Tomita. M. (1992) Identification of the

bactericidal domain of lactoferrin. Biochim. Biophys Acta

1121. 130-136.

[I41 Lindh, E. (1975) Increased resistance of immunoylobulin A

dimers to proteolytic degradation after hinding of aetrerory

component. J. Immunol. 114, 2X4-286.