Chiral Resolution and Molecular ModelingInvestigation of rac-2-Cyclopentylthio-6-

[1-(2,6-difluorophenyl)ethyl]-3,4-dihydro-5-methylpyrimidin-4(3H)-one (MC-1047), a Potent

Anti-HIV-1 Reverse Transcriptase Agent of theDABO Class

M.G. QUAGLIA,1* A. MAI,1 G. SBARDELLA,1 M. ARTICO,1 R. RAGNO,2 S. MASSA,3 D. DEL PIANO,4G. SETZU,4 S. DORATIOTTO,4 AND V. COTICHINI5

1Istituto Pasteur-Fondazione Cenci Bolognetti, Dipartimento di Studi Farmaceutici, Universita degli Studi diRoma “La Sapienza,” Roma, Italy

2Dipartimento di Studi di Chimica e Tecnologie delle Sostanze Biologicamente Attive, Universita degli Studidi Roma “La Sapienza,” Roma, Italy

3Dipartimento Farmaco Chimico Tecnologico, Universita degli Studi di Siena, Siena, Italy4Dipartimento di Biologia Sperimentale, Universita degli Studi di Cagliari, Cittadella Universitaria,

Cagliari, Italy5Istituto Superiore di Sanita, Roma, Italy

ABSTRACT rac-2-Cyclopentylthio-6-[1-(2,6-difluorophenyl)ethyl]-3,4-dihydro-5-methylpyrimidin-4(3H)-one (MC-1047) is a potent inhibitor of HIV-1 multiplication inacutely infected cells. MC-1047 racemate has been resolved by chiral HPLC using, aschiral stationary phase (CSP), a commercially available (R,R)-Whelk-01 column. Theoptical purity and the circular dichroism (CD) of the two resolved enantiomers weredetermined and their biological activities tested in in vitro assays. Molecular modelinginspection of the binding of (R) and (S) enantiomers to the non-nucleoside binding site(NNBS) of reverse transcriptase (RT) using the defined model of F2-S-DABO/RTcomplex indicates the (R) enantiomer as the more active isomer. Chirality 13:75–80,2001. © 2001 Wiley-Liss, Inc.

KEY WORDS: enantiomeric separation; binding mode; non-nucleoside reverse tran-scriptase inhibitor; chiral stationary phase; circular dichroism

Anti-HIV-1 dihydro-alkoxy-benzyl-oxopyrimidines(DABOs)1–3 and the related thioanalogues (S-DABOs)4,5

are uracil and thymine derivatives having as peculiar sub-stituents an alkyl(cycloalkyl)oxy or alkyl(cycloalkyl)thiomoiety at position 2 and a phenylmethyl group at position6 of the pyrimidine ring. Replacement of phenylmethylwith the bulkier 1-naphthylmethyl moiety gave DATNOs,6a potent subfamily of DABOs. Further improvement of an-tiviral potency was obtained by introducing at the C-6 po-sition a 2,6-difluorophenylmethyl chain (difluorothio-DABOs, F2-S-DABOs).7

Molecular modeling investigations of F2-S-DABOs led toidentification of a predictive reverse transcriptase/difluoroDABO (RT/F2-S-DABO) complex model, which in-dicated that introduction of small alkyl groups at the meth-ylene bridge placed between benzene and pyrimidine ringswould modulate the “butterfly-like” spatial arrangement soas to increase antiviral potency. Therefore, we synthesizeda series of 5-alkyl-2-alkylthio(or cycloalkylthio)-6-[1-(2,6-dihalophenyl)ethyl]-3,4-dihydropyrimidin-4(3H)-one de-rivatives and tested them in vitro against HIV-1 in MT-4

cells. Among the newly synthesized derivatives, rac-2-cyclopentylthio-6-[1-(2,6-difluorophenyl)ethyl]-3,4-dihydro-5-methylpyrimidin-4(3H)-one (MC-1047) was found to ex-ert inhibitory activity against RT in the nanomolar range(Table 1).

MC-1047 was prepared (Fig. 1) by reacting the precur-sor 6-[1-(2,6-difluorophenyl)ethyl]-3,4-dihydro-2-thioxo-pyrimidin-4(3H)-one with cyclopentyl bromide in the pres-ence of potassium carbonate. The synthesis of the precur-sor thioxopyrimidinone was performed starting from ethyl4-(2,6-difluorophenyl)-3-oxopentanoate, as reported else-where.8 MC-1047 contains a chiral carbon atom and there-fore exists as a racemic mixture. Due to its high potency,we decided to attempt to separate the related enantiomers.In fact, it is well established that the enantiomers of variouschiral drugs share substantial differences in pharmacoki-

*Correspondence to: Prof. M. Giovanna Quaglia, Dipartimento di StudiFarmaceutici, Universita degli Studi di Roma “La Sapienza,” P.le A. Moro5, I-00185 Roma, Italy. E-mail: quaglia@uniroma1.itReceived for publication 9 March 2000; Accepted 13 July 2000

CHIRALITY 13:75–80 (2001)

© 2001 Wiley-Liss, Inc.

netic, pharmacodynamic, and toxic properties.9,10 There-fore, the chiral discrimination of a racemate allows deter-mination of the pharmacological activity of each enantio-mer and to safely use new chiral drugs. The study of therelationships between stereochemistry and biological ac-tivity of a chiral drug requires a versatile and accuratemethod for the enantioseparation to allow the separate as-say of the activity of the racemate and both (+) and (−)enantiomers.

In this article, we describe the enantioseparation of MC-1047, the characterization of each enantiomer by UV andCD spectra, and the pharmacological tests used to checktheir biological activity. Furthermore, molecular modelingof the binding mode of (R) and (S) enantiomers of MC-1047 to the HIV-1 RT NNBS links the greatest biologicalpotency with the (R) enantiomer.

EXPERIMENTALChemistry

Melting points were determined on a Buchi 530 meltingpoint apparatus and are uncorrected. Infrared (IR) spectra(KBr pellets) were recorded on a Perkin-Elmer 297 instru-ment. 1H NMR spectra were recorded at 200 MHz on aBruker AC 200 spectrometer. Chemical shifts are reportedin d (ppm) units relative to the internal reference tetra-methylsilane (Me4Si). Reactions were routinely checkedby TLC and 1H NMR. TLC used aluminum-backed silicagel plates (Merck DC-Alufolien Kieselgel 60 F254). Devel-oped plates were visualized under UV light. Solvents werereagent grade. Concentration of solution after reaction in-volved the use of a rotary evaporator operating at approxi-mately 0.02 Torr. Organic solutions were dried over anhy-

drous sodium sulphate. Analytical results for C, H, N, F,and S were to within ±0.40% of the theoretical values.

rac-2-Cyclopentylthio-6-[1-(2,6-difluorophenyl)ethyl]-3,4-dihydro-5-methylpyrimidin-4(3H)-one

(rac-MC-1047)A mixture of rac-6-[1-(2,6-difluorophenyl)ethyl]-3,4-

dihydro-5-methyl-2-thioxopyrimidin-4(3H)-one8 (0.48 g, 1.7mmol), cyclopentyl bromide (0.27 g, 0.19 ml, 1.8 mmol),and potassium carbonate (0.23 g, 1.7 mmol) in 2 ml ofanhydrous N,N-dimethylformamide was stirred at roomtemperature for 8 h. The reaction content was poured oncold water (100 ml) and extracted with ethyl acetate (3 × 50ml). The organic layers were collected, washed with brine(3 × 50 ml), dried overnight, then evaporated to furnishcrude rac-MC-1047 as a solid, which was purified by crys-tallization. C18H20F2N2OS, molecular weight 350.43. Re-crystallized from cyclohexane; mp 196–197°C; yield%: 60.IR (KBr): 2900 (NH), 1640 (CO) cm−1. 1H NMR (CDCl3) d1.43–1.67 (m, 9H, C2,3,4,5-H cyclopentyl and CHCH3), 1.99–2.17 (m, 5H, C2,5-H cyclopentyl and C5-CH3), 3.98–4.02 (m,1H, C1-H cyclopentyl), 4.55–4.59 (q, 1H, CHCH3), 6.77–6.85(m, 2H, C3,5-H Ar), 7.14–7.22 (m, 1H, C4-H Ar), 11.88 (bs,1H, NH exchangeable with D2O) ppm.

InstrumentationThe enantiomeric separation was carried out with a Wa-

ters 600E pump equipped with Waters 996 photodiode-array detector and a Jasco 1595 circular dichroism detector(CD), aided by a Millennium 32 work station (Waters, Mil-ford, MA, USA).

Reagents and ChemicalsAll solvents and chemicals used were of HPLC or ana-

lytical grade (Merck, Darmstadt, Germany).

Chromatographic AnalysisThe enantioseparation of rac-MC-1047 was carried out

using a commercial (R,R)-Whelk-01 column (244 mm × 4mm i.d., 5 mm, Merck) as CSP. n-Hexane:2-propanol (90:10v/v) was used as mobile phase at flow rate of 0.8 ml/min.The UV spectra and CD signs of separated enantiomerswere checked. Afterwards, each isomer was separately col-lected from several runs by injecting 10 µg of racemate onthe same analytical column. Each fraction was divided intwo parts and then separately dried under vacuum. Thetwo aliquots were used to determine the optical purity ofthe separated enantiomers, using the same CSP, and toevaluate their pharmacological activity, respectively.

Antiviral Assay ProceduresCompounds. MC-1047 and the separated enantiomers

were dissolved in DMSO (200 mM) and then diluted intoculture medium. These solutions were used to carry outthe antiviral assays.

Cells and viruses. MT-4, C8166, and H9/IIIB cellswere grown at 37°C in a 5% CO2 atmosphere in RPMI 1640medium, supplemented with 10% fetal calf serum (FCS),100 IU/ml penicillin G, and 100 µg/ml streptomycin. Cellcultures were checked periodically for the absence of my-coplasma contamination with a MycoTect Kit (Gibco,

TABLE 1. Cytotoxicity and anti-HIV-1 activity ofMC-1047 enantiomersa

Compound

µM

SIeCCb50 ECc

50 ICd50

rac-MC-1047 >200 0.005 0.005 >40,000[+]-MC-1047f >200 0.002 0.008 >100,000[−]-MC-1047 >200 0.7 1.0 >286

aData represent mean values for three separate experiments. Variationamong triplicate samples was less than 10%.bCompound dose required to reduce the viability of mock-infected cells by50%, as determined by the MTT method.cCompound dose required to achieve 50% protection of MT-4 cells fromHIV-1-induced cytopathogenicity, as determined by the MTT method.dCompound dose required to inhibit the HIV-1 rRT activity by 50%.eSelectivity index, CC50/EC50.f[+] and [−] are referred to CDC values recorded at 244 nm.

Fig. 1. Synthesis of rac-MC-1047.

76 QUAGLIA ET AL.

Grand Island, NY). Human immunodeficiency virus type-1(HIV-1, IIIB strain) was obtained from supernatants of per-sistently infected H9/IIIB cells. HIV-1 solutions had titersof 4.5 × 106 cell culture ID50 (CCID50)/ml.

HIV titration. Titration of HIV-1 was performed inC8166 cells by the standard limiting dilution method (di-lution 1:2, four replica wells per dilution) in 96-well plates.The infectious virus titer was determined by light micro-scope scoring of cytopathicity after 4 days of incubation;the virus titers were expressed as CCID50/ml.

Anti-HIV assays. The compound activity test againstHIV-1 multiplication in acutely infected cells was based onthe inhibition of virus-induced cytopathicity in MT-4 cells.Fifty µl of culture medium containing 1 × 104 cells wereadded to each well of flat-bottom microtiter trays contain-ing 50 µl of culture medium with or without test com-pounds at different concentrations. Then 20 µl of an HIVsuspension containing 100 (HIV-1) CCID50 were added.After 4 days of incubation at 37°C, the number of viablecells was determined by the MTT method.11 Compoundcytotoxicity was evaluated in parallel with antiviral activity,based on the viability of mock-infected cells as monitoredby the MTT method.

RT assays. Assays were performed as previously de-scribed.7 Purified rRT was assayed for its RNA-dependentpolymerase-associated activity in a 50 µl volume contain-ing: 50 mM Tris-HCl (pH 7.8), 80 mM KCl, 6 mM MgCl2,1 mM DTT, 0.1 mg ml−1 BSA, 0.5 OD260 unit ml−1 tem-plate:primer [poly(rC)-oligo(dG)12-18] and 10 mM [3H]-dGTP (1 Ci mmol−1). After incubation for 30 min at 37°C,samples were spotted on glass fiber filters (Whatman GF/A), and the acid-insoluble radioactivity was determined.

Molecular Modeling

Molecular modeling studies were performed with thesoftware packages SYBYL 6.512 and MACROMODEL 6.513

running on a Silicon Graphics O2 R10000 225 MHz. Thestructures of MC-1047 enantiomers were modeled from thecrystal structure of a representative F2-DABO (unpub-lished results) using the SYBYL fragment library. The ini-tial conformations for the docking experiments wereachieved by superimposing the uracil rings of (R)- and(S)-MC-1047 upon that of TNK-651, previously extractedfrom its RT complex (PDB entry code 1rt2).14 The geom-etry of the NNBS of HIV-1 RT was taken from the structureof HIV-1 RT/TNK-651 complex. All the residues within 20Å from any ligand atom (TNK-651) were used to define theNNBS. The starting structures of the complexes RT/(R)-and RT/(S)-MC-1047 were obtained by replacing TNK-651with the two enantiomers inside the NNBS, by the super-imposition of the uracil ring atoms. The docking of the twoDABO derivatives was carried as follows: first, to investi-gate the NNBS flexibility a SYBYL/FLEXIDOCK proce-dure was performed allowing the rotation of all the rotat-able ligand torsion angles and the sidechain torsion anglesof a 4 Å residue core in the NNBS; second, a Monte Carlosearch was carried out with the MACROMODEL package(AMBER force field). The four torsion angles of DABOcompounds were rotated randomly allowing an 8 Å core ofatoms of the NNBS to relax during the minimization. Anexternal fixed shell of 10 Å was also included for the long-range interactions. After 1,000 Monte Carlo steps (MCMMroutine) the global minimum was found four times and,hence, used to describe the binding mode of (R)- and (S)-MC-1047. The strain energies of the receptor and the li-gand were obtained by the difference of the steric energiesof the bound conformation (from a single point calculation)and the unbound molecule after a MACROMODEL mini-mization to its nearest local minimum. For the ligands thestrain energies were calculated in water using the con-tinuum solvation model of Gibbs (GBSA),15,16 while for thereceptor the calculations were performed in vacuum. Inthis latter case the 137 water molecules found in the HIV-1

Fig. 2. Chromatographic enantioresolution of rac-MC-1047.

MOLECULAR MODELING OF MC-1047 77

RT/TNK-651 complex were retained in all the calculationsto allow inclusion of some solvation effect.

RESULTS AND DISCUSSION

The enantioseparation of MC-1047 was made on theWhelk-01 chiral stationary phase (CSP). This column, a

highly enantioselective CSP of wide applicability, is(3S,4R)-[4-(3,5-dinitrobenzamido)-tetrahydrophenantrene],covalently bonded to 3-propyl silica.17 This separates race-mic compounds by taking advantage of p–p electron inter-actions and hydrogen bonding and it can be used for theenantioseparation of compounds possessing p-donor or

Fig. 3. CD spectra of MC-1047 enantiomers recorded at 244 nm (a) and at 290 nm (b).

Fig. 4. Dose-response curves of rac-MC-1047, [+]-MC-1047, [−]-MC-1047, and nevirapine.

78 QUAGLIA ET AL.

p-acceptor groups. Thus, enantiomers with either p-acidicor p-basic groups can be separated efficiently.

MC-1047 bears in its structure an electron-acceptor con-jugated p-system located near the stereogenic center, soWhelk-01 seemed to be suitable for the enantioseparationof such a racemate. As expected, using Whelk-01 columnwith normal phase elution (n-hexane/2-propanol) a goodenantioseparation of MC-1047 racemate was achieved (Fig.2). The obtained good separation allowed us to separatelycollect the individual enantiomers of MC-1047, used forchiroptical and biological assays.

The analyses were carried out with an HPLC instrumentwith UV and CD detectors on line which allowed the iden-

tification of the two separated enantiomers and to evaluatetheir chiroptical behavior. The CD spectra, recorded by astop flow system during HPLC analysis, showed differ-ences in the Cotton effect when the spectrum was recordedat 244 nm (Fig. 3a) or at 290 nm (Fig. 3b). The substituentsin the 4-pyrimidinone ring gave variations in the n-p* tran-sitions at different wavelengths.18,19 Therefore, the CDspectra of the first eluted enantiomer showed a relativelyhigh positive band in the 240 nm region and a lower nega-tive one in the 300 nm region. To have a higher sensitivitythe CD spectra were recorded at 244 nm and the firsteluted enantiomer was named [+]-MC-1047 and the secondeluted one [-]-MC-1047. Obviously, if the spectrum is re-

Figure 5. The two enantiomers of MC-1047 (R in green, S in red) into the HIV-1 RT NNBS.

MOLECULAR MODELING OF MC-1047 79

corded in the 300 nm region (Fig. 3b) the first eluted en-antiomer should be [-] while the second one [+]. The goodresolution obtained (a = 1.47) between two enantiomersallowed separate collection of them by more than 100 runsusing the analytical chiral column. The optical purity ofeach collected enantiomer was tested using the same CSP.The enantiomeric excess was greater than 97%.

The individual enantiomers collected were used for theantiviral assays. The antiviral activity of rac-MC-1047 andeach isomer was evaluated in vitro, as described in theExperimental section. The data obtained, summarized inTable 1, show that [+]-MC-1047 has a higher activity withrespect to [-]-MC-1047 and rac-MC-1047. The dose–response curves of racemate, [+]-MC-1047 and [-]-MC-1047are reported in Figure 4 in comparison with nevirapine, acommercially available NNRTI.

The greater activity of [+]-MC-1047 suggests that thisenantiomer is a potential candidate for use as a drug. Fur-ther investigations of the activity against a panel of resis-tant mutants and the toxicity of [+]-MC-1047 are beingpublished elsewhere.8

Molecular Modeling Investigation

The inspection of the binding mode model of difluoro-DABOs into the NNBS of the HIV-1 RT (unpublished data)leads us to indicate the (R) form as the most active be-tween the two enantiomers. In fact, the (R) methyl of MC-1047, together with the C5 methyl, could act as promotersof the “trigger action” on Tyr181, which is forced to makea clockwise rotation around the ca-cb linkage of about 120°.

The theoretical binding mode of the (R) and (S) MC-1047 enantiomers to the NNBS is shown in Figure 5. The(R)-methyl of MC-1047 points toward an empty lipophilicpocket created by the sidechains of Lys103, Val106, andVal179, thus providing further steric interactions inside theNNBS, whereas the (S)-methyl, which turns in the oppo-site direction, does not produce new effective interactionswith the NNBS. Moreover, in the (S) enantiomer of MC-1047 the chiral methyl lies in an eclipsed configurationcompared with the C5 methyl, thus forcing the molecule toassume, in aqueous medium, a conformation not suitablefor initial binding. On the contrary, the (R) enantiomerdoes not suffer any peculiar intramolecular interaction.

These observations are supported by the comparison ofboth the ligand and receptor strain energies. (The receptorstrain energy is the difference of steric energy between thebound and the first local minimum of free molecule (inwater for the ligand and in vacuum for the receptor).) (R)-MC-1047, starting from the bound conformation in water,frees 23.79 KJ/mol and its corresponding NNBS needs168.35 KJ/mol to relax, while the (S) enantiomer spends27.93 KJ/mol to adopt the binding conformation, and thecorresponding binding site requires 179.89 KJ/mol to hostit (Fig. 5). These observations indicate that the (R) enan-tiomer of MC-1047 is the most active isomer, since it needslower energy to form the respective complex with HIV-1 RT.

LITERATURE CITED

1. Artico M, Massa S, Mai A, Marongiu ME, Piras G, Tramontano E, LaColla P. 3,4-Dihydro-2-alkoxy-6-benzyl-4-oxopyrimidines (DABOs): a

new class of specific inhibitors of human immunodeficiency virus type1. Antiviral Chem Chemother 1993;4:361–368.

2. Tramontano E, Marongiu ME, De Montis A, Loi AG, Artico M, MassaS, Mai A, La Colla P. Characterization of the anti-HIV-1 activity of3,4-dihydro-2-alkoxy-6-benzyl-4-oxopyrimidines (DABOs), new non-nucleoside reverse transcriptase inhibitors. Microbiologica 1994;17:269–279.

3. Massa S, Mai A, Artico M, Sbardella G, Tramontano E, Loi AG, ScanoP, La Colla P. Synthesis and antiviral activity of new 3,4-dihydro-2-alkoxy-6-benzyl-4-oxopyrimidines (DABOs), specific inhibitors of hu-man immunodeficiency virus type-1. Antiviral Chem Chemother 1995;6:1–8.

4. Mai A, Artico M, Sbardella G, Massa S, Loi AG, Tramontano E, ScanoP, La Colla P. Synthesis and anti-HIV-1 activity of thio analogues ofdihydroalkoxybenzyloxopyrimidines. J Med Chem 1995;38:3258–3263.

5. Ettorre A, Mai A, Artico M, Massa S, De Montis A, La Colla P. 6-(3-Methylbenzyl)-2-(2-methylpropyl)thio-4(3H)-pyrimidinone (DABO622). Acta Cryst 1996;C52:2115–2117.

6. Mai A, Artico M, Sbardella G, Quartarone S, Massa S, Loi AG, DeMontis A, Scintu F, Putzolu M, La Colla P. Dihydro(alkylthio)(naph-thylmethyl)oxopyrimidines: novel non-nucleoside reverse trasncrip-tase inhibitors of the S-DABO series. J Med Chem 1997;40:1447–1454.

7. Mai A, Artico M, Sbardella G, Massa S, Novellino E, Greco G, Loi AG,Tramontano E, Marongiu ME, La Colla P. 5-Alkyl-2-(alkylthio)-6-(2,6-dihalophenylmethyl)-3,4-dihydropyrimidin-4(3H)-ones: novel potentand selective dihydro-alkoxy-benzyl-oxopyrimidine derivatives. J MedChem 1999;42:619–627.

8. Mai A, Sbardella G, Artico M, Massa S, Ragno R, Ettorre A, NovellinoE, Greco G, Loi AG, Tramontano E, Marongiu ME, La Colla P. Bindingmode and structure based drug design of novel NNRTI 2-alkylthio-6-[1-(2,6-difluorophenyl)alkyl]-3,4-dihydro-5-alkylpyrimidin-4(3H)-ones.J Med Chem (in press).

9. Ariens EJ, Soudijin W, Timmerman PBMWM. Stereochemistry andbiological activity of drugs. Oxford: Blackwell; 1983.

10. Wamer JW. Drug stereochemistry. New York: Marcel Dekker; 1993.

11. Pauwels R, Balzarini J, Baba M, Snoeck R, Schols D, Herdewijn P,Desmyster J, De Clercq E. Rapid and automated tetrazolium-basedassay for the detection of anti-HIV compounds. J Virol Methods 1988;20:309–321.

12. SYBYL molecular modeling system, version 6.5. St. Louis: TRIPOSAssociated.

13. Mohamadi F, Richards NGJ, Guida WC, Liskamp R, Lipton M, CaufieldC, Chang G, Hendrickson T, Still WC. MacroModel — an integratedsoftware system for modeling organic and bioorganic molecules usingmolecular mechanics. J Comput Chem 1990;11:440–467.

14. Hopkins AL, Ren J, Esnouf RM, Willcox BE, Jones EY, Ross C, Miya-saka T, Walker RT, Tanaka H, Stammers DK, Stuart DI. Complexes ofHIV-1 reverse transcriptase with inhibitors of the HEPT series revealconformational changes relevant to the design of potent non-nucleoside inhibitors. J Med Chem 1996;39:1589–1600.

15. Still WC, Tempczyk A, Hawley RC, Hendrickson T. Semianalyticaltreatment of solvation for molecular mechanics and dynamics. J AmChem Soc 1990;112:6127–6129.

16. Qiu D, Shenkin PS, Hollinger FP, Still WC. The GB/SA continuummodel for solvation. A fast analytical method for the calculation ofapproximate born radii. J Phys Chem A 1997;101:3005–3014.

17. Quaglia MG, Desideri N, Bossu E, Sgro R, Conti C. Enatioseparationand anti-rhinovirus activity of 3-benzylchroman-4-ones. Chirality 1999;11:495–500.

18. Lightner DA. Analytical applications of circular dichroism. Purdie andBritain, editors. Oxford: Elsevier Sciences; 1994. 131–173 pp.

19. Wellman KM, Briggs WS, Djerassi C. Optical rotatory dispersion stud-ies. C. Variable-temperature circular dichroism studies of ring-conformational and rotational equilibria in cyclohexanones. J AmChem Soc 1965;87:73–81.

80 QUAGLIA ET AL.