Separation of Peptides from Myoglobin Enzymatic Digests by RPLC

Gradients on the Retention and Separation of Peptides from a Cytochrome-c Digest by Reversed-Phase Liquid Chromatography. L. A. Riddle , G. Guioch...
0 downloads 0 Views 119KB Size
Anal. Chem. 2005, 77, 3425-3430

Articles

Separation of Peptides from Myoglobin Enzymatic Digests by RPLC. Influence of the Mobile-Phase Composition and the Pressure on the Retention and Separation Nicola Marchetti† and Georges Guiochon*

Department of Chemistry, The University of Tennessee, Knoxville, Tennessee 37996-1600, and Division of Chemical and Analytical Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831

The influence of the mobile-phase composition and the pressure on the chromatographic separation of the peptides from the enzymatic digest of myoglobin was studied under linear conditions. The retention behavior of these tryptic peptides was measured under isocratic conditions with different mobile-phase compositions, ranging from 9 to 28% (v/v) acetonitrile in 0.1% (v/v) aqueous trifluoroacetic acid. The effect of the pressure was studied by analyzing the separation of the tryptic peptides under different average column pressures between 14 and 220 bar, at 13, 20, and 26% (v/v) acetonitrile. The differences between the partial molar volumes of these peptides in the stationary and mobile phases were derived from these results. All the measurements were performed on a 10-cm-long C18-bonded, end-capped monolithic column. The results obtained illustrate the highly complicated behavior of the complex peptide mixtures afforded by tryptic digestion. The capacity factors of the analyzed peptides do not depend linearly on the acetonitrile concentration but follow exactly a quadratic relationship. The adsorption changes of partial molar volumes are in good agreement with other literature data. The consequences of the influence of the average column pressure (hence of the flow rate) on the column phase ratio and on the retention factors of the peptides are discussed. The retention pattern of the complex mixture is affected by both the mobile-phase composition and the pressure, and the resolution of certain peptide pairs is so much affected by the pressure that inversions in the elution order of some pairs are observed. The influence of the pressure on the chromatographic behavior of analytes has always been the object of benign neglect by analysts, if not of utter ignorance. This neglect is understandable as long as the separations of mixtures of small molecules are * To whom correspondence should be addressed. Tel.: +1-865-9740-733. Fax: +1-865-9742-667. E-mail: [email protected]. † On leave from the Department of Chemistry, University of Ferrara, via L. Borsari 46, I-44100 Ferrara, Italy. 10.1021/ac050541c CCC: $30.25 Published on Web 04/29/2005

© 2005 American Chemical Society

considered.1-4 The influence of the pressure on the retention of analytes is essentially governed by Le Chatelier’s principle.2,5,6 This principle states that an increase of the pressure applied will shift the equilibrium toward the composition or the state that has the lower specific volume. Basic thermodynamics shows that the change in the Gibbs free energy of the equilibrium of an analyte between the two phases of a chromatographic system is proportional to the difference between the partial molar volumes of this compound in the two phases of the system.2,4,5 Intuition suggests and experimental measurements confirm that the magnitude of this change increases with increasing molecular weight of organic molecules. Accordingly, the effect of the pressure on the retention of low molecular weight compounds is small enough to be negligible in practice.3,4,7 Unfortunately, this is no longer true for compounds having large molecules and particularly for peptides and proteins for which the effect can become considerable.4,8-10 The retention factor of insulin doubles when the average column pressure increases by 200 bar. However, this effect is comparable to that of a change in the organic modifier concentration of a few percent.10 When closely related macromolecules (for which the differences between their partial molar volumes in the liquid and solid phases are also close) are separated, the effect is limited to a slight shift in the retention of the whole chromatogram, with minor changes in its resolution. The effect is significant when the mixture contains components that are closely eluted but have markedly different molecular volumes. If very high pressures are applied, as is now possible, the chromatogram may be dramatically affected by a change in flow (1) McGuffin, V. L.; Evans, C. E. J. Microcolumn Sep. 1991, 3, 513. (2) Guiochon, G.; Sepaniak, M. J. J. Chromatogr. 1992, 606, 148. (3) McGuffin, V. L.; Chen, S. J. Chromatogr. 1997, 69, 35. (4) Szabelski, P.; Cavazzini, A.; Kaczmarski, K.; Liu, X.; Horn, J. V.; Guiochon, G. J. Chromatogr., A 2002, 950, 41. (5) Martire, D. E.; Locke, D. C. Anal. Chem. 1967, 39, 921. (6) Bidlingmeyer, B. A.; Rogers, L. B. Sep. Sci. 1972, 7, 131. (7) McGuffin, V. L.; Chen, S. Anal. Chem. 1997, 69, 930. (8) Bylina, A.; Ulanowicz, M. Anal. Chem. (Warsaw) 1998, 43, 955. (9) Chen, S.; Ho, C.; Hsiao, K.; Chen, J. J. Chromatogr. 2000, 891, 207. (10) Liu, X.; Zhou, D.; Szabelski, P.; Guiochon, G. Anal. Chem. 2003, 75, 3999.

Analytical Chemistry, Vol. 77, No. 11, June 1, 2005 3425

rate that affects the pressure profile along the column.11,12 All the parameters that are usually considered as constant in chromatography (e.g., mobile-phase density, packing porosity, column hold-up volume, diffusion coefficients) become pressure dependent,12 and the theoretical framework accounting for the thermodynamic effects of pressure becomes complex.13 Thus, it is important to investigate in detail the influence of the pressure on the retention of peptides and proteins that constitute the class of analytes of greatest current importance. In this work, we investigate the combined influence of the mobile-phase composition and the pressure on the separation of enzymatic digests of myoglobin by RPLC. Peptide maps or fingerprints of proteolyzed proteins are usually obtained by RPLC because this method affords great flexibility in the choice of the experimental parameters and permits marked enhancements of the resolution of the digest mixture by adjusting the different factors that influence the thermodynamic and kinetic processes involved in the interaction of the peptides and the nonpolar stationary phase. The solvophobic theory shows that the isocratic retention factor is related to the temperature, the pressure, the mobile-phase composition, the molar volume of the solvent, and the phase ratio (see later, Theory section).14-18 Hence, we also report in this work a study concerning the influence of mobilephase composition on the tryptic peptides retention. Besides the fundamental importance of this work, its results will be useful since peptide mapping is most important in proteomics analysis.19 The interpretation of the chromatograms obtained when using high-performance, high-pressure instruments may often be more difficult than that of more conventional separations. THEORY Influence of the Organic Modifier on the Retention. Nonpolar interactions between a solute like a protein or peptide and a hydrophobic matrix can be described using the solvophobic theory, which states that the isocratic retention factor, k′, can be related to the overall difference between the Gibbs free energies of the solute dissolved in the mobile phase and adsorbed on the stationary phase. However, this model is difficult to use in practice, so empirical and semitheoretical models have been developed to describe or to attempt to predict the dependence of the retention factors on the mobile-phase composition. One of the most important of these models was suggested by Snyder et al.20-22 It introduces the concept of linear solvent strength (LSS). According (11) Martin, M.; Blu, G.; Guiochon, G. J. Chromatogr. Sci. 1973, 11, 641. (12) Martin, M.; Guiochon, G. J. Chromatogr., A. In press. (13) Martire, D. E.; Boehm, R. E. J. Phys. Chem. 1987, 91, 2433. (14) Boldingh, J. Experientia 1948, 4, 270. (15) Sinanoglu, O. In Molecular Association in Biology; Pullman, B., Ed.; Academic Press: New York, 1968. (16) Sinanoglu, O. Theor. Chim. Acta 1974, 33, 279. (17) Horvath, C.; Melander, W.; Molnar, I. J. Chromatogr. 1976, 125, 129. (18) Melander, W. R.; Corradini, D.; Horvath, C. J. Chromatogr. 1984, 317, 67. (19) Blackstock, W., Mann, M., Eds. Proteomics: a Trend Guide; Elsevier: London, 2000. (20) Snyder, L. R.; Stadalius, M. A. High-Performance Liquid Chromatography: Advances and Perspectives; Academic Press: New York, 1986; Vol. 4. (21) Snyder, L. R.; Kirkland, J. J. Introduction to Modern Liquid Chromatography; Wiley Interscience: New York, 1979. (22) Jandera, P.; Churacek, J. Gradient elution in column liquid chromatography: theory and practice; Journal of Chromatography Library 31; Elsevier: Amsterdam, 1985.

3426

Analytical Chemistry, Vol. 77, No. 11, June 1, 2005

to this model, the solute retention factor, k′, can be expressed as a linear function of the volume fraction of the organic modifier in the mobile phase, φ, as23

ln k′ ) ln k0′- Sφ

(1)

where k′0 is the retention factor extrapolated at φ ) 0 and S is a constant related to the nature of the organic modifier and to the structure of the solute and the stationary phase. The LSS model gives a good approximation of the experimental results for binary mobile phases over a limited range of capacity factors or mobilephase compositions. When the range of variation is larger, a quadratic equation

ln k′ ) ln k′0 - S1φ + S2φ2

(2)

is more appropriate to account for the experimental data. Volume Change of the Solute upon Adsorption. The influence of pressure on the equilibrium constants and the retention factors can be assessed by considering the classical relationships of phase equilibrium in thermodynamics. The equilibrium constant of adsorption, K, is

K ) Cs/Cm

(3)

where Cm and Cs are the concentrations of the solute in the mobile and the stationary phases at equilibrium, respectively. The retention factor is the product of the equilibrium constant and the column phase ratio (k′ ) FK). The change in the Gibbs free energy, ∆G, of this equilibrium is given by10

∆G ) -RT ln K

(4)

The influence of the pressure and the temperature on the equilibrium constant are obtained by differentiation of this equation. By differentiating eq 4 with respect to P we obtain

(∂ln∂PK)

T

)

1 ∂∆G RT ∂P

( )

T

)

∆Vm ∆Vk′ + ∆Vφ ) RT RT

(5)

where ∆Vm is the difference between the partial molar volumes of the solute molecule in the mobile and the stationary phases, and ∆Vk′ and ∆VF are the contributions to ∆Vm of the pressure dependences of ln k′ and ln F, respectively. EXPERIMENTAL SECTION Equipment. A HP 1100 liquid chromatography system (Agilent Technologies, Palo Alto, CA) was used for all the experimental determinations. This instrument is equipped with a multisolvent delivery system, a vacuum degasser, a variable-wavelength detector, a high-pressure flow cell, and a computer data station. Analitycal injections were performed through a Rheodyne 7725i manual injection valve (Cotati, CA), using a 20-µL sample loop. A capillary restrictor (a polyether ether ketone (PEEK) tube, i.d. ) 0.0025 in, Upchurch, Oak Harbor, WA) cut to the desired length, (23) Snyder, L. R. High-Performance Liquid Chromatography: Advances and Perspectives; Academic Press: New York, 1980; Vol. 1.

was placed at the UV detector outlet, to allow an easy independent adjustment of the average column pressure (ACP) and the flow rate.4,10 Mobile Phase and Chemicals. The mobile phase was a solution of acetonitrile, water, and trifluoroacetic acid (TFA). Two mother solutions were used: 0.1% (v/v) TFA in H2O (pump channel A) and 0.1% (v/v) TFA in ACN (pump channel B). All solvents were filtered before use on a hydrophilic modified poly(vinylidene fluoride) Durapore membrane, 0.22-µm pore size (Millipore, Billerica, MA). Acetonitrile was HPLC grade from Fischer Scientific (Fair Lawn, NJ). Water was Milli-Q grade (Millipore Corp., Billerica, MA). TFA was purchased from Aldrich (Milwaukee, WI). Thiourea (Fluka, Milwaukee, WI) was chosen as the unretained tracer to measure the column hold-up volume. Myoglobin, trypsin, and Tris were purchased from Sigma (St. Louis, MO). Column. A Chromolith Performance RP-18e, 100 mm × 4.6 mm, column was used. This C18-bonded, end capped monolithic column (column 35) was one of the lot of six columns used by Gritti and Guiochon,24 a gift from Merck (Darmstadt, Germany). According to the manufacturer, the average size of the pores in the bare porous silica is 130 Å for the mesopores and 1.95 µm for the macropores. The values of the column hold-up volume given later are the average of the results of three consecutive injections of thiourea, made after column equilibration at the given mobilephase composition. Thiourea is very slightly retained on alkylbonded silica (k′ < 0.05).25 Enzymatic Digestion. The digestion of myoglobin was carried out with trypsin (1 mg/mL in 1 mM HCl). No reducing and alkylating steps were used in this procedure. Myoglobin was solubilized in 40 mM Tris-HCl buffer solution (pH 8.5). The amount of trypsin to add was calculated in order to have a final ratio of 1:50 (w/w) between enzyme and substrate. The digestion was performed at 37 °C for 24 h and then stopped with a 10 mM HCl solution. A reference solution, containing Tris-HCl buffer and the same amount of trypsin as the previous digestion solution, was left at the same temperature and time condition to evaluate the extent of possible autodigest processes. Measurements of Experimental Data. All the experimental data were obtained at 1.0 mL/min mobile phase flow rate. The ACP is the average of the pressures at the column inlet and outlet. It was varied within the interval 14-220 bar. Injections of thiourea into the instrument without the column, which is replaced with a zero-volume connector, were performed to measure the extracolumn volume of the instrument. The extracolumn volume is 0.09 mL. Similar injections of thiourea into the instrument with the column permit the measurement of the column hold-up volume, equal to the difference between the retention time of thiourea and the extracolumn volume. The column hold-up volume was found in the range 1.452-1.537 (13% ACN v/v), 1.434-1.507 (20% ACN v/v), and 1.418-1.466 mL (26% ACN v/v), depending on the pressure. The phase ratio F is the ratio of the volumes occupied by the mobile and the stationary phases in the column, F ) t/(1 - t), where t is the total column porosity. This porosity is the fraction of the volume of the column tube available to the liquid phase, t (24) Gritti, F.; Guiochon, G. J. Chromatogr., A 2003, 1021, 25. (25) Gritti, F.; Guiochon, G. J. Chromatogr., A 2005, 1070, 1.

Figure 1. Separation of myoglobin digest under gradient elution and an average column pressure of 18 bar. Mobile phase: 0.1% (v/ v) TFA in H2O (pump channel A) and 0.1% (v/v) TFA in ACN (pump channel B). Gradient: linear 5-40% (v/v) ACN-0.1% (v/v) TFA. Gradient time: 30 min.

) V0/Vcol (with Vcol ) πdc2/Lc ) 1.662 mL). The phase ratio depends on the mobile-phase composition, due to the variation of the degree of swelling of the bonded layer with the concentration of the organic modifier. Due to the compressibility of the liquid phase, of the bonded layer, and of silica, and to the elasticity of the tube, both column and column hold-up volumes depend on the pressure. The phase ratios at different pressures and mobile-phase compositions were derived from the total porosity measured under these different conditions. The peak profiles were recorded with the UV detector, the signal being monitored at 210 nm. Each injection (thiourea and digest mixture) was repeated three times consecutively. The reproducibility of the elution times of the unretained compound and of all the tryptic peptides was high, with a relative error on these times always lower than 0.5 and 1%, respectively. The temperature at which the data were collected was 20 °C. RESULTS AND DISCUSSION It is important to underline that all the data measured and reported here were acquired under linear chromatography conditions. Figures 1 and 2 show two chromatograms of the peptide digest obtained in gradient elution, with the same program for the mobile-phase composition but under two different ACP conditions. The differences illustrate the significant influence of pressure on the retention pattern. The arrows indicate the peaks that were shifted by the pressure increase. However, all the measurements discussed here were made under isocratic conditions. Influence of the Mobile-Phase Composition on the Retention Factors. The influence of the mobile-phase composition on the retention behavior of the tryptic digest peptides of myoglobin was studied first. Figure 3 reports all the data acquired on the evolution of the retention factors of 14 selected fragments when the acetonitrile concentration in the mobile phase was incresed from 9 to 28% (Figure 3, symbols). Looking at the experimental data, it is evident that for most tryptic fragments, especially for those monitored in a wide range of ACN concentration (e.g., fragments 7, 9, and 10), the trend is not linear (Figure 3, dashed Analytical Chemistry, Vol. 77, No. 11, June 1, 2005

3427

Table 1. Changes in Molar Volume and Retention Factor of Some Myoglobin Tryptic Peptides under Different Concentrations of Acetonitrile

tryptic peptide 1 2 4 5 6 7 8

tryptic peptide Figure 2. Separation of myoglobin digest under gradient elution and an average column pressure of 162 bar. Mobile phase and gradient conditions are the same as in Figure 1.

8 7 9 10 11 12

tryptic peptide 10 12 13 14

Figure 3. Logarithmic capacity factors for 14 selected tryptic fragments from myoglobin digest against the volume fraction of organic solvent in water-acetonitrile-0.1 TFA isocratic mobile phases.

lines). This is highlighted by the accordance between the quadratic fitting functions and the overlying experimental points (see eq 2). This behavior, which has often been observed for small molecules, has already been reported for small peptides and proteins. This result confirms that the LSS model (see Theory section) cannot be applied in a broad range of mobile-phase compositions. The lines shown in Figure 3 were obtained by fitting the experimental data to a quadratic function of the ACN concentration. The agreement between the experimental data and this second-order relationship between ln k′ and φ is excellent and suitable for an accurate description of the retention of all the fragments in the whole range analyzed. It is interesting to note that many lines cross over, especially those corresponding to all the peptides between 4 and 8. As a consequence, some peptides that are closely eluted, yet well resolved at relative low ACN concentrations (e.g., 2 and 3, 11 and 12, or 13 and 14), are no longer resolved at higher ACN concentrations. A change of the ACN concentration by 3-4% is sufficient to turn a pair of well-resolved peaks into a multiplet. This rapid variation of the retention and separation factors with the ACN concentration illustrates the importance of a careful 3428 Analytical Chemistry, Vol. 77, No. 11, June 1, 2005

13% ACN (v/v) retention factor (k′) Pav ∼ 14 bar Pav ∼ 221 bar 0.21 0.32 1.21 1.35 2.38 2.74 6.19

0.21 0.35 1.43 1.60 2.96 3.33 7.94

20% ACN (v/v) retention factor (k′) Pav ∼ 15 bar Pav ∼ 220 bar 0.21 0.43 0.58 2.31 5.87 11.32

0.28 0.51 0.73 2.81 7.20 15.53

26% ACN (v/v) retention factor (k′) Pav ∼ 15 bar Pav ∼ 211 bar 0.35 0.92 4.43 5.02

0.43 1.26 6.14 8.02

∆Vm (mL mol-1) -50.11 -60.05 -69.64 -69.20 -74.62 -71.76 -76.61

∆Vm (mL mol-1) -71.02 -59.67 -66.80 -62.43 -63.28 -76.66

∆Vm (mL mol-1) -55.16 -69.14 -71.71 -89.76

optimization of the parameters of separations performed in gradient elution. Influence of the Pressure on the Retention and Separation Factors. On the basis of the data in Figure 3, we chose three different isocratic conditions and measured the pressure dependence of the retention factors under these conditions, from which we calculated the adsorption change of the partial molar volume of some of the tryptic peptides (see Table 1). In Figure 4 the values of of ln k′ are plotted versus the ACP for four of these peptides. The figure shows the data measured for fragments 7 and 8 at 13 and 20% ACN (v/v). Note that the elution order of these two fragments is inverted when the ACN concentration is raised from 13 to 20% (v/v) while their separation factor (R ) k′II/k′I, where k′II > k′I, see Table 2, second and third columns) is barely affected by the pressure. The slopes of the lines in Figure 4 show that the contribution ∆Vk′ (see eq 5) to the overall variation of ∆Vm is -6.7% for peptide 7 and +15.6% for peptide 8 (see Table 2). This explains why the isocratic separation factor of the two peptides changes in opposite directions, increasing by 5.0% at 13% ACN when the pressure is increased from 14 to 221 bar while it decreases by 11.2% at 20% ACN. The column hold-up volume changes with the solvent strength and with the average column pressure. The total porosity, t, increases; hence the phase ratio decreases with increasing hold-up volume and with increasing ACP (see Influence of the Organic Modifier on the Retention). Since the relationship between ln F and P is linear, we can derive from the plot in Figure 5 the slope (∆VF) and calculate ∆VF. The analysis of the experimental data shows that this contribution to ∆Vm is predomi-

Figure 4. Pressure dependence of ln k′ in the case of tryptic peptides 7, 8, 10, and 12 measured at different mobile-phase composition (symbols): fragment 7 with 13 (]) and 20% ACN (v/v) ([); fragment 8 with 13 (0) and 20% ACN (v/v) (9); fragment 10 with 20 (4) and 26% ACN (v/v) (2); fragment 12 with 20 (O) and 26% ACN (v/v) (b). Dashed lines represent linear regression fits to experimental points.

Figure 5. Phase ratio versus the average column pressure at different acetonitrile concetrations in water.

Table 2. Contribution ∆Vk′ to the Overall ∆Vm, Calculated from Line Slopes of the Plot Reported in Figure 4 for Two Different Peptides and Their Separation Factors

tryptic peptide 7 8

tryptic peptide 8 7 10 12

tryptic peptide 10 12

13% ACN (v/v) separation factor (R) Pav ∼ 14 bar Pav ∼ 221 bar 2.26

2.38

20% ACN (v/v) separation factor (R) Pav ∼ 15 bar Pav ∼ 220 bar 2.05

1.82

4.90

5.53

26% ACN (v/v) separation factor (R) Pav ∼ 15 bar Pav ∼ 211 bar 2.63

2.93

∆Vm (mL mol-1) -22.27 -27.14

∆Vm (mL mol-1) -32.18 -20.79 -23.11 -37.50

∆Vm (mL mol-1) -25.03 -39.02

nant for both mobile-phase compositions (see Figure 5), with values of -49.56 mL/mol at 13% ACN and -38.96 mL/mol at 20% ACN. This means that the isobaric separation factor increases by ∼ +21.4% and explains why the overall value of ∆Vm (Table 1) increases only by +7.3% for peptide 8 but by +16.8% for peptide 7. Figure 4 shows the data for two more retained fragments, 10 and 12, at higher ACN concentrations, 20 and 26% (v/v). In this case, the elution order does not change with the mobile-phase composition (see Figure 3). The two peptides are well resolved in the whole composition range. There is an interference between peptides 11 and 12, which coelute at 26% ACN, which is why fragment 11 is not listed in Table 1, in the 26% ACN section. Calculations made from the experimental data set acquired for these peptides show a similar negative variation of ∆Vm for both

Figure 6. Band profiles of tryptic fragments 11 and 12 recorded at different average column pressures.

peptides when [ACN] increases from 20 to 26%, -11.6% for fragment 10 and -9.8 for 12. This is because the ∆Vk′ contributions (see Table 2) do not change with the modifier concentration. The separation factors of the two peptides decrease by about -46.3% at the lowest pressure and by -47% at the highest pressure (Table 2). The phase ratio contribution to ∆Vm is -30.14 mL/mol at 26% ACN and changes by +22.6% when [ACN] increases from 20 to 26%, as in the previous case. Finally, the importance of the influence of pressure on the resolution between peptides 11 and 12 is illustrated in Figure 6. This figure shows that an increase of the average column pressure can affect considerably a chromatogram. CONCLUSION This work illustrates the changes in the retention pattern of a complex mixture with changes in the mobile-phase composition and the column pressure. The logarithms of the capacity factors of the different peptide fragments do not vary linearly with the acetonitrile concentration but do follow a quadratic relationship. The parameters of the relationship vary considerably from one peptide to the next. This explains rapid inversion of the elution order and suggests that the structures of the different peptides are poorly related. The complexity of the differential migration Analytical Chemistry, Vol. 77, No. 11, June 1, 2005

3429

pattern makes the separation of these mixtures most difficult to achieve even with gradient elution. The results obtained regarding the influence of pressure on the retention are in excellent agreement with similar results available for peptides and small proteins. The values of ∆Vm obtained for the different tryptic peptides of myoglobin are between -50 and -90 mL/mol, comparable to -100 mL/mol obtained for insulin. This distribution is wide enough and the effect of pressure sufficiently important for the retention times to vary significantly with the pressure and for the resolution of certain pairs to depend on the pressure. Fragments that coelute under 14 bar are well resolved under 210 bar. TFA is widely used in the analysis of peptides and proteins because it is a superior eluent modifier. It acts as an

3430

Analytical Chemistry, Vol. 77, No. 11, June 1, 2005

ion-pairing agent with peptides and significantly improves peak profiles, column efficiency, and band resolution. The influence of the pressure on the retention of peptides may be related to the influence of pressure on these ion-pairing interactions of TFA. ACKNOWLEDGMENT This work was supported in part by Grant CHE-02-44693 of the National Science Foundation and by the cooperative agreement between the University of Tennessee and the Oak Ridge National Laboratory. Received for review March 30, 2005. Accepted March 31, 2005. AC050541C