acquisition of ordered conformation by the n-terminal domain of the human small proline rich 2...

AD

EP*T

R

tdtsrl2CschodtdnSpCe

t

ttd

CTP

rrrTcem

Biochemical and Biophysical Research Communications 262, 395–400 (1999)

Article ID bbrc.1999.1215, available online at http://www.idealibrary.com on

cquisition of Ordered Conformation by the N-terminalomain of the Human Small Proline Rich 2 Protein

leonora Candi,* Gerry Melino,* Marco Sette,† Sergio Oddi,*ietro Guerrieri,* and Maurizio Paci†,1

Laboratory of Biochemistry, Istituto Dermopatico dell’Immacolata, c/o Department of Experimental Medicine, Universityor Vergata, Rome, Italy; and †Department of Chemical Science and Technology, University Tor Vergata, Rome, Italy

eceived July 16, 1999

epithelial surface by desquamation, but these layers ofdlmtbklbnoaopcto

hfmotvsosrtaifeoe

fetp

The cornified cell envelope (CE) is a crucial struc-ure for barrier function in terminally differentiatedead stratified squamous epithelia. It is assembled byransglutaminase enzymes (TGases) that cross-linkeveral proteins such as loricrin and the small prolineich (SPR) proteins. Human SPR2 protein is cross-inked with widely differing efficiencies by TGases 1,, and 3 using exclusively residues in the N- and-terminal domains. In order to understand if the ab-ence of the cross-linking catalyzed by TGases in theentral domain is due to the conformation adopted, weave investigated the structural properties in solutionf three peptides that correspond to the N-terminalomain, to three repeats of the central domain, and tohe C-terminal domain. Together, the NMR and CData strongly indicate the presence of a highly flexibleon a-helix, non b-sheet structure in SPR2. Thus,PR2 appears to function as a flexible cross-bridgingrotein to provide tensile strength or rigidity to theE of the stratified squamous epithelia in which it isxpressed. © 1999 Academic Press

Key Words: barrier function; small proline-rich pro-ein; cornified envelope; nuclear magnetic resonance.

Stratified squamous epithelia undergo a complexerminal differentiation program which has many fea-ures typical of terminal differentiation followed by celleath (1, 2). The dead cells are eventually lost from the

1 To whom correspondence should be addressed at Department ofhemical Science and Technology, University of Tor Vergata, Via dior Vergata 135, 00133, Rome, Italy. Fax: 139-06-7259-4328. E-mail:[email protected] used: CE, cornified envelope; SPR, small proline

ich (protein); TGase, transglutaminase; NMR, nuclear magneticesonance; TFE, 2,2,2-trifluoroethanol; GdHCl, guanidinium chlo-ide; CD, circular dichroism; NMR, 1H-nuclear magnetic resonance;OCSY, total correlation spectroscopy; ROESY, rotating-frame nu-lear Overhauser effect spectroscopy; NOESY, nuclear Overhauserffect spectroscopy; NOE, nuclear Overhauser effect; RMSD, rootean square deviation.

395

ead epithelial cells function largely to prevent wateross and act as a physical barrier against the environ-

ent. The cornified cell envelope (CE) is a 10–15 nmhick layer of highly insoluble protein which is assem-led just beneath the plasma membrane in livingeratinocytes by disulfide bonds and Ne-(g-glutamyl)-ysine or N1,N8-bis(g-glutamyl)spermidine isopeptideonds formed by the catalytic action of transglutami-ase (TGase) enzymes. TGases form irreversible intra-r intermolecular isopeptide bonds between glutaminend lysine residues. The structural protein precursorsf the CE are: loricrin, keratin, desmoplakin, envo-lakin, periplakin, involucrin, cystatin a, elafin, tri-hohyalin, repetin, pancornulins, calcium binding pro-eins such as annexin I, and various selected membersf the SPR family.The SPR group is the most complex, consisting in

uman and mouse of 11-12 members divided into threeamilies, SPR1 (two members), SPR2 (eight to eleven

embers), and SPR3 (one member). We have previ-usly shown that only glutamine and lysine residues inhe head and tail domains are used in cross-linking inivo and in vitro (3, 4, 5, 6, 7). This implies that SPRserve as cross-bridging proteins by joining themselvesr other proteins such as loricrin, involucrin, desmo-omal proteins, etc., by the use of multiple adjacentesidues in the end domains. Furthermore, a correla-ion between the amount of SPR proteins used in CEsnd the presumed physical requirements for mechan-cal strength and toughness of the epithelium has beenound. Thus, the SPRs serve as biomechanical modifi-rs of the physical properties of the CE structure inrder to fulfill the particular requirements of differentpithelia to withstand physical trauma (5).Recent biochemical experiments have shown that,

or the SPR1 (3) and SPR2 (7) proteins, each TGasenzyme preferentially crosslinks certain specific glu-amine and lysine residues with high specificity, im-lying that multiple enzymes are required to cross-link

0006-291X/99 $30.00Copyright © 1999 by Academic PressAll rights of reproduction in any form reserved.

them in vivo. Moreover, in the case of SPR1 proteins,tpstp

assrptwttSti

tcSrptctstp

M

c(wwc(ppppsp

wm5twTrmspd(d

odwam

R

TvtlSespbrtaisdtpbfathra

sAos

Vol. 262, No. 2, 1999 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

he data imply an obligatory temporal order to thisrocess: first cross-linking by the cytosolic TGase 3 intohort oligomers, which in turn are later cross-linked byhe membrane-associated TGase 1 enzyme into largeolymers into the CE barrier.All members are built according to a common plan of

mino (head) and carboxy (tail) domains containingeveral adjacent glutamine and lysine residues, whichpan a central domain composed of a series of peptideepeats of 8–9 residues which are rich in prolines. Therecise sequence of these repeats allow distinction intohe three families. The numbers of repeats variesidely both between members of the family and be-

ween species: for example, human SPR1 proteins con-ain 6 repeats while mouse contain 13 or 14; all humanPR2 members contain 3 repeats, while mouse pro-eins contain 3.5–9 repeats; and the number of repeatsn SPR3 varies from about 16–30 in different species.

The studies reported here on the conformational fea-ures of the N-terminal, C-terminal and three perfectlyonserved repeats of the central domain of humanPR2 protein revealed a strong general tendency toemain unstructured even in the presence of the veryowerful structuring medium trifluoroethanol, leadingo the conclusion that this feature confers to the un-rosslinked portion a marked flexibility. Among thehree peptides, the N-terminal domain of SPR2 washown to be more structured than the other two pep-ides corresponding to the C-terminal and the threeerfectly conserved repeats of the central domains.

ATERIALS AND METHODS

Circular dichroism. The circular dichroism spectra were re-orded on a Jasco J-710 circular dichroism spectropolarimeterHachigi City, Japan). The spectrum was recorded from 200–250 nmith readings every 0.2 nm at 25°C. Variable-temperature spectraere recorded at 30°C (unless indicated) by keeping the temperature

onstant with a heat bath. Peptides were synthesized by IDI-IRCCSRome, Italy). The peptide concentration was 100 mM of HPLC-urified peptide in 10 mM phosphate buffer at pH 5.3. For the SPR2rotein spectrum a concentration of 0.4 mg/ml was used and therotein was dialysed in phosphate buffer solution at pH 7.0. A 0.1 cmath length quartz cuvet was used to record all the spectra. Theolvent spectrum was subtracted from that of SPR2 peptide androtein and a total of 4 scans were accumulated for each sample.

NMR experiments and sequential assignments. All the spectraere recorded in a Bruker 400 MHz AM instrument at 298°K with 3M peptide concentration in degassed 10 mM phosphate buffer (pH

.3), 15 mM ditiothreitol to prevent sulphidryl group oxidation. Allwo dimensional data were acquired in the phase-sensitive modeith the saturation of the water signal during the relaxation delay.OCSY data were collected with the data matrix (t1 5 512, t2 5 2 K);elaxation delay 5 1.41 s; number of transients 5 64; isotropicixing 5 0.068 s, 0.098 s; MLEV-17 pulse sequence was used for the

pin-lock. NOESY (8) data was collected with similar acquisitionarameters and for 0.280 and 0.300 s of mixing times. ROESY (9, 10)ata was collected in the phase-sensitive mode. Data matrix wast1 5 512; t2 5 2 K); mixing time applied was 0.3 s; relaxationelay 5 1.5 s; number of transients 5 64. The sequential assignment

396

f peptide A of SPR-2 is obtained by combining TOCSY and ROESYata at 298 K and ROESY at 305 K and 278 K. ROESY experimentsere also performed in different TFE concentrations with the samecquisition parameters (10, 30, and 70% v/v). One TOCSY experi-ent was performed in 10% TFE with an isotropic mixing of 0.080 s.

ESULTS AND DISCUSSION

Secondary structure by CD and NMR spectroscopy.he biochemical data already reported (3, 11, 7) show aery different behaviour of lysine residues present inhe head and tail domains in comparison with theysines present in the central domain repeats of thePR proteins. In order to evaluate the structural prop-rties of these three domains, and to identify a possibletructure-function relationship, we performed an ex-erimental evaluation of the conformational featuresy circular dichroism (CD) and 1H-nuclear magneticesonance (NMR) on these three separate peptides ofhe SPR2 protein. The sequence of the three peptidesre reported in Fig. 1. For the central repeats domaint was decided that the three central repeats, whichhowed an identical sequence, would be studied. Theominant structural feature in the human SPR2 pro-ein and in the three peptides is predicted to be theresence of a b-turn conformation (4.00 3 1024 proba-ility). These repetitive b-turns are the only predictedeature for the proline-rich central repeat domain. Thebsence of predicted a-helical or b-sheet structures inhe repeated sequences can be accounted for by theigh proline content. A minimum of five favourableesidues are required to form an a-helix and six for

b-sheet, proline destabilizing either structure. A

FIG. 1. Complete amino acid sequence of hSPR2. Underlinedequences correspond to the peptides studied by CD and NMR.mino acids are shown in a single letter code. The percentage yieldf Gln and Lys residues utilization is shown below the amino acidequence: upper row, TGase1 enzyme; lower row, TGase3 enzyme.

pieTtCtc

m2apctprats

a

uecsttbdtoebtsttdps

as

tst


olyproline II structure is also unlikely to form, requir-ng consecutive proline residues or proline residues atvery third position along the polypetide chain (12, 13).hese predictions for the central repeats domain struc-ure of hSPR2 have been confirmed in hSPR3. In fact,D and NMR studies on a peptide, also corresponding

o three repeats of the central domain of hSPR3, indi-ate the presence of a class B b-turn per repeat (11).

Circular dichroism spectroscopy. CD spectra of hu-an SPR2 peptides show a very similar profile (Fig.

A, 2B, and 2C). The different value of molar ellipticityt 198 nm for SPR2 peptides compared to that of therotein is probably caused by the different length of thehain (14, 15). The far-UV spectrum is dominated byransitions associated with the amide groups of theeptide backbone. The most striking feature in thisegion is the absence of shapes characteristic of eithern a-helical- or b-sheet-rich conformation, supportinghe prediction based on structural analysis. Indeed, thepectrum most closely resembles that of a random coil.Upon increase of temperature, only slight changes

re visible in the CD spectra reported above. In partic-

FIG. 2. CD spectra of hSPR2 peptides at different temperature to the N-terminal domain of hSPR2; (B) spectra of the peptide correspectra of the peptide corresponding to the C-terminal domain of hSPemperature transition of 20–30–40–60°C.

397

lar all the three peptides show a marked increase ofllipticity around 222 nm. Particularly, the peptideorresponding to the N-terminal domain (Fig. 2A)hows a marked change centered at 225 nm indicatinghe formation of a small amount of a-helical conforma-ion. Much less pronounced are the effects visible inoth the three central repeats and in the C-terminalomain peptides (Fig. 2B and 2C, respectively). Thehermal behaviour, the reversibility and the presencef an isodichroic point at 212 nm suggest that onequilibrium with at least two conformers is detectabley CD. Higher temperatures shift the equilibrium be-ween the conformers. This can be explained as a re-ult of the stabilization by hydrophobic interactions ofhe secondary structures which are favoured at highemperatures (16). The temperature could produceenaturation or a change in the conformation of theroline peptide-bond from trans to cis as previouslyuggested (17).Increasing the concentration of 2,2,2-Trifluoroeth-

nol (TFE), from 10% to 70% (v/v) in water resulted inimilar changes in the far- and near-UV regions as

sitions (20–30–40–60°C). (A) Spectra of the peptide correspondingnding to the identical 3 repeats of the central domain of hSPR2; (C)(D) difference spectra of hSPR2 peptides at 225 nm obtained during

ranpoR2;

t3whtdlsdrateia

tass2cwbp

cb

taNcstsemsetwtpsi

Nat

poo


hose observed with increasing temperature (Fig. 3A,B, and 3C). TFE is a structure inducing solvent,hich enhances intramolecular interactions such asydrogen-bonding and electrostatic salt bridges in pro-eins and peptides. In this case also, we observe aecrease in intensity and a shift in the peak wave-ength from 200 to 203–204 nm, with an increase of thehoulder at 225 nm (Fig. 3A). Both the experimentsone with the peptides corresponding to the centralepeats and C-terminal domains respectively, (Fig. 3Bnd 3C) show an isodichroic point (212 nm), indicatinghe presence of a conformational equilibrium. How-ver, the dichroic profile indicates only small changesn conformation with no overall transition toward-helix or beta conformation of the peptides.The differences observed in the CD spectrum ob-

ained both by increasing the temperature (Fig. 2D)nd the concentration of TFE (Fig. 3D) at 225 nm, corre-pond to a class-B b-turn spectrum according to the clas-ification scheme of Woody (18), with a minimum around25 nm and a maximum in the region 200–205 nm. Thislass-B spectrum is the most common class associatedith any b-turn conformation (19, 20). The increase in-turn conformation content with both increasing tem-erature and hydrophobicity of the solvent would indi-

FIG. 3. Effect of TFE addition on the CD spectra of the hSPR2 peptide corresponding to the N-terminal domain of hSPR2; (B) spectf hSPR2; (C) spectra of peptide corresponding to the C-terminal dbtained using TFE concentration of 0, 10, 30, and 70%.

398

ate that the conformation rich in b-turn minimizesoth hydrophobic interactions and hydrogen-bonding.

NMR spectroscopy. The NMR spectroscopy washus applied in order to better investigate the second-ry structural elements seen in the CD spectra of the-terminal end of SPR2. The total correlation spectros-

opy (TOCSY) and rotating-frame overhauser effectpectroscopy (ROESY) were used to obtain the sequen-ial assignment of the protons belonging to the con-tituent amino acids. ROESY data were also used tostimate inter-proton distances. In water, the two di-ensional NMR studies on SPR2 N-terminal peptide

how a substantially extended conformation of the mol-cule. The internal dynamics of the molecule revealhat the effective correlation time for most of the pair-ise interproton interactions are fast enough to elude

he observation of the nuclear overhauser effect cross-eaks by NOESY experiments. However, in water-TFEolution several inter-residue interactions have beendentified in the ROESY spectra.

Table I shows the individual assignments of theMR spectrum for the N-terminal peptide achieved bycquiring TOCSY and ROESY at different tempera-ures and at different concentrations of deuterated

tides using TFE concentration of 0, 10, 30, and 70%. (A) Spectra ofof the peptide corresponding to the 3 repeats of the central domainain of hSPR2; (D) difference spectra of hSPR2 peptides at 225 nm

epraom

Trtbotettta1tfli

stefsetssSfcndcbl

N-terminal peptide of SPR2, the more structured inTt

cthcmuFtdmammml

arfl

pprNpn

TABLE I

R


FE in order to remove the overlapping due to theepeated sequences. The spin systems were connectedhrough interesidue Hai-NHi 1 1 ROEs. This methodreaks down at proline residues because of the absencef the backbone amide proton, a severe limitation inhe assignment of these proline-rich peptides. How-ver, the sequential Ha(i)-ProHd(i 1 1) ROE connec-ivities were useful to identify the sequential connec-ions. The strong sequential NOEs Ha(i)-NH(i 1 1) inhe tract between Ser1 and Gln4 and between Cys7nd Gln9 and, conversely, the absence of NH(i)-NH(i 1) in the tract between Gln4 and Cys7 indicate a par-ially structured conformation in these two tracts and aexible conformation in the central region around res-

dues 5 and 6.

Local structure of the repeated sequence. Due to themall number of observed NOEs, the calculations ofhree-dimensional structures can not rule out the pres-nce of less populated conformers and a possible con-ormational exchange in the intermediate NMR timecale. The limited number of NOEs found in ROESYxperiments (Fig. 4A and 4B) in TFE did not allow uso apply the usual protocol to determine the structureolution by distance-geometry and molecular dynamicsimulation. In fact, the results recently obtained withPR3 where a higher number of distance constraints

rom NOEs than in SPR2 were found and used in thealculation only as upper limits have led to a largeumber of conformational families (11). This is clearlyue to the lack of long-range NOE connectivities. In thease of SPR3 structure, the comparison by RMSD forackbone atom coordinates showed that only a veryimited convergence was obtained.

Those observations led to the conclusion that also the

Human SPR2 N-terminal Peptide

esidued NH(ppm)

d Ha(ppm)

d Hb(ppm)

d Hg(ppm)

Other(ppm)

Ser-1 — 4,16 3,974,03 (b9)

— —

Tyr-2 8,59 4,48 3,023,05 (b9)

7,08 6,80 (e)

Gln-3 8,36 4,11 1,98 2,44 —Gln-4 7,87 4,17 2,06

2,03 (b9)2,43

2,34 (g9)—

Gln-5 8,19 4,16 1,94 2,38 7,29 (z)Gln-6 7,87 4,26 2,02 2,50

2,46 (g9)Cys-7 7,86 4,45 2,07 — —Lys-8 7,90 4,34 1,86

1,77 (b9)1,44

1,42 (g9)1,68 (d)2,99 (e)7,51 (z)

Gln-9 8,02 4,23 2,142,05 (b9)

2,362,40 (g9)

7,24 (z)

Pro-10 — 4,37 2,312,28 (b9)

2,032,01 (g9)

3,68 (d)3,76 (d9)

399

FE, is a rather flexible portion of the molecule andhere are only very partially rigid local regions.

In vivo consequences of the structural data. Bio-hemical studies on the recombinant human SPR2 pro-ein revealed that only the Glu and Lys residues on theead and tail domain sequences are involved inrosslinking in vitro by TGases (7), in excellent agree-ent with the in vivo studies (6, 21, 22, 23). In partic-lar, TGase1 utilizes only some residues, as shown inig. 1. One probable explanation of the utilization ofhese residues could come out from our CD and NMRata. This structural data suggests that, although hu-an SPR2 protein has a low order of organized second-

ry structure, the SPR2 repeats are more flexible andobile in comparison with the N-terminal head do-ain and the C-terminal tail domains. This flexibilityay prevent the TGases from recognizing the residues

ocalized in the repeats as substrates.Together the present data on human SPR2, as well

s the previous data for SPR1 and SPR3 proteins,eveal that the members of this family possess veryexible structures with a low order of organized sec-

FIG. 4. (A) ROESY spectrum (at 300 ms, 298°K) of 3 mM of theeptide corresponding to the N-terminal domain of hSPR2 in phos-hate buffer (pH 5.3) containing 15 mM ditiothreitol, fingerprintegion. Ha-NH sequential connectivities are indicated by lines. (B)OE contacts found in ROESY spectra of the hSPR2 N-terminaleptide obtained as reported under Materials and Methods; thick-ess indicate the intensity of NOEs observed.

ondary structure. These results confirm our cur-rflpt

A

NgMtt

R

8. Jeener, J., Meier, B. H., Bachmann, P., and Ernst, R. R. (1979)

1

1

1

1

1

1

11

1

1

2

2

2

2


ent model which suggests that SPRs may serve asexible cross-bridging ‘spacers’ between structuralroteins that constitute the protein backbone ofhe CE.

CKNOWLEDGMENTS

Fabio Bertocchi is gratefully acknowledged for the execution of theMR experiments. The work was partially carried out thanks torants from Min. San., MURST 40% 1997 and 1998; NationalURST Project “Structural Biology;” and the target project Bio-

echnology of CNR, Telethon E.872 to G.M. and Telethon 417/bio E.C.

EFERENCES

1. Polakowska, R. R., and Haake, A. R. (1994) Cell Death Differ. 1,11–18.

2. Eckert, R. L., and Welter, J. F. (1996) Cell Death Differ. 3,373–383.

3. Candi, E., Tarcsa, E., Idler, W. W., Kartasova, T., Marekov,L. N., and Steinert, P. M. (1999) J. Biol. Chem. 274, 7226–7237.

4. Steinert, P. M. (1995) Cell Death Differ. 2, 33–40.5. Steinert, P. M., Candi, E., Kartasova, T., and Marekov, L. N.

(1998a) J. Struct. Biol. 122, 76–85.6. Steinert, P. M., Kartasova, T., and Marekov, L. N. (1998b)

J. Biol. Chem. 273, 11758–11769.7. Tarcsa, E., Candi, E., Kartasova, T., Idler, W. W., Marekov,

L. N., and Steinert, P. M. (1998) J. Biol. Chem. 273, 23297–23303.

400

J. Chem. Phys. 71, 4546–4553.9. Bax, A., and Davis, D. G. (1985) J. Magn. Res. 63, 207–213.0. Griesinger, C., and Ernst, R. R. (1987) J. Magn. Reson. 75,

261–271.1. Steinert, P. M., Candi, E., Tarcsa, E., Marekov, L. N., Sette, M.,

Paci, M., Ciani, B., Guerrieri, P., and Melino, G. (1999) CellDeath Differ., in press.

2. Isemura, T., Asakura, J., Shibata, S., Iasemura, S., Saitoh, E.,and Sanada, K. (1983) Int. J. Peptide Protein Res. 21, 281–287.

3. Shibata, S., Asakura, J., Isemura, T., Isemura, S., Saitoh, E., andSanada, K. (1984) Int. J. Peptide Protein Res. 23, 158–165.

4. Raj, P. A., Edgerton, M., Levine, M. J., and Salivary (1990)J. Biol. Chem. 265, 3898–3905.

5. Raj, P. A., Johnsson, M., Levine, M. J., Nancollas, G. H. (1992)J. Biol. Chem. 267, 5968–5976.

6. Tanford, C. (1970) Adv. Prot. Chem. 24, 1–95.7. Darrell Fontenot, J., Tiandra, N., Bu, D., Ho, C., Montelaro,

R. C., and Finn, O. J. (1993) Cancer Research 53, 5386–5394.8. Woody, R. W. (1976) in Peptides, Polypeptides, and Proteins

(Blout, E. R., Ed.), pp. 338–350, Wiley, New York.9. Smith, J. A., and Pease, L. G. (1980) CRC Crit. Rev. Bioc. 8,

314–399.0. Tatham, A. S., Drake, A. F., and Shewry, P. R. (1985) Biochem.

J. 226, 557–562.1. Steinert, P. M., and Marekov, L. N. (1995) J. Biol. Chem. 270,

17702–17711.2. Steinert, P. M., and Marekov, L. N. (1997) J. Biol. Chem. 272,

2021–2030.3. Robinson, N. A., Lapic, S., Welter, J. F., and Eckert, R. L. (1997)

J. Biol. Chem. 272, 12035–12046.

acquisition of ordered conformation by the n-terminal domain of the human small proline rich 2...

Documents