Exploring Amino Acid Side Chain Decomposition Using Enzymatic

Aug 19, 2009 - Spencer S. Walse, Michael J. Plewa and William A. Mitch* ... John D. Sivey , Stanley C. Howell , Doyle J. Bean , Daniel L. McCurry , Wi...
0 downloads 0 Views 728KB Size
Download PDF
Anal. Chem. 2009, 81, 7650–7659

Exploring Amino Acid Side Chain Decomposition Using Enzymatic Digestion and HPLC-MS: Combined Lysine Transformations in Chlorinated Waters Spencer S. Walse,† Michael J. Plewa,‡ and William A. Mitch*,† Department of Chemical Engineering, Environmental Engineering Program, Yale University, Mason Lab 313b, 9 Hillhouse Avenue, New Haven, Connecticut 06520-8286, and Department of Crop Sciences, College of Agricultural, Consumer, and Environmental Sciences, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801 Characterizing the transformations of polypeptides is important across a broad range of scientific disciplines. As polypeptides are an important constituent of dissolved organic matter within seawater and freshwater, it is important to understand their (bio)geochemical fate. Oxidants, formed in blood as part of the immunological response or applied to waters for disinfection, react with polypeptides to form transformation products that may exert toxicity. An analytical method was developed to characterize and quantify modifications to the side chains of amino acid residues within polypeptides. Enzymatic digestion of polypeptides using Pronase E, a protease cocktail, proved preferable to common strong acid digestion techniques, because the circumneutral pH conditions employed during enzymatic digestion prevent artifacts arising from extreme pH conditions. Lysine nitrile, one of the predicted transformation products of lysine residues within polypeptides, was destroyed during strong acid digestion but not enzymatic digestion. Due to the potential variability in enzymatic digestion efficiencies, the liberation of a mass-labeled leucine monomer from an octapeptide spiked standard was employed as a measure of complete digestion efficiency for each sample and enabled quantification of modified amino acid residues within polypeptides. A multivariate statistical analysis was conducted to evaluate the influence on digestion efficiency of Pronase E loadings, salinity, natural organic matter concentration, and pH across the range of conditions relevant to blood, seawater, and concentrated freshwaters and disinfected drinking/recreational waters. At Pronase E loadings of 10 mg, the analysis indicated that digestion efficiencies ranged from 25 to 55% over the range of conditions expected for typical drinking waters concentrated from 1 L to 10 mL. The analytical method was applied to triplicate 1 L samples of a chlorinated tap water and a chlorinated indoor pool water. For the tap water, the digestion efficiency was 47.2% ((11.1% relative standard deviation), and the lysine nitrile concentration was 104.6 ng/L ((6.8 ng/L standard deviation). For the pool water, the digestion efficiency was 23% ((15% relative standard deviation); although the occurrence of lysine nitrile was verified, matrix effects must be overcome for quantification. * Corresponding author. Tel: (203) 432-4386. Fax: (203) 432-4387. E-mail: [email protected]. † Yale University. ‡ University of Illinois at Urbana-Champaign.

7650

Analytical Chemistry, Vol. 81, No. 18, September 15, 2009

Understanding the decomposition of polypeptides is important across a broad range of scientific disciplines. Protein is an important constituent of biological fluids, such as blood. Polypeptides in drinking waters may be increasing as the decreasing availability of pristine water supplies fostered by population growth is encouraging utilities to exploit waters impaired by algal blooms or wastewater effluents.1 Both of these waters feature relatively fresh biopolymers shed as either algal or bacterial exudates and higher dissolved organic nitrogen (DON) concentrations reflective of elevated polypeptide content.2 Similarly, recreational waters feature biopolymer loads from bathers, released via urine, sweat, and dermal shedding.3 As the main source of organic carbon to the ocean is primary biological production, it is not surprising that, together with polysaccharides, polypeptides dominate the 20-40% of dissolved organic matter (DOM) that has been characterized.4,5 Although many processes contribute to the decomposition of polypeptides within these systems, the side chains of amino acids are important reaction foci for abiotic transformations.6,7 The aromatic side chains of tyrosine and tryptophan may be susceptible to photolysis, while the electron-rich thiol and amine groups in the side chains of certain amino acids, such as methionine, lysine, and histidine, render these moieties primary targets for reactions with oxidants.6,7 In addition to microbial transformations,8,9 oceanic DOM is processed by photolysis and reactive oxygen species (e.g., hydroxyl radicals).10 Characterizing side chain modifications would clarify how these processes either break (1) Desalination and Water Purification Technology Roadmap: A Report of the Executive Committee; Program Report #95; Desalination & Water Purification Research & Development; U.S. Bureau of Reclamation: Washington, DC, 2003. (2) Westerhoff, P.; Mash, H. J. Water Supply: Res. Technol.-Aqua 2002, 51 (8), 415–448. (3) Tricker, A. R.; Pfundstein, B.; Preussmann, R. Carcinogenesis 1992, 13, 563–568. (4) Ogawa, H.; Tanoue, E. J. Oceanogr. 2003, 59, 129–147. (5) McCarthy, M.; Pratum, T.; Hedges, J.; Benner, R. Nature 1997, 390, 150– 153. (6) Hawkins, C. L.; Pattison, D. I.; Davies, M. J. Amino Acids 2003, 25, 259– 274. (7) Pattison, D. I.; Davies, M. J. Chem. Res. Toxicol. 2001, 14 (10), 1453– 1465. (8) Hollibaugh, J. T.; Azam, F. Limnol. Oceanogr. 1983, 28 (6), 1104–1116. (9) McCarthy, M. D.; Benner, R.; Lee, C.; Hedges, J. I.; Fogel, M. L. Mar. Chem. 2004, 92, 123–134. (10) Blough, N. V.; Zepp, R. G. In Active Oxygen in Chemistry; Foote, C. S.; Valentine, J. S.; Greenberg, A.; Liebman, J. F., Eds.; Blackie Academic & Professional: London, 1995; 280-333. 10.1021/ac901064u CCC: $40.75  2009 American Chemical Society Published on Web 08/19/2009

down polypeptides to low molecular weight, bioavailable products or convert them to poorly characterized refractory materials featuring turnover times of up to ∼4000 years.4 While oxidants such as hypochlorous acid are used to disinfect drinking waters and recreational waters, their reactions with dissolved organic materials form a range of chemical byproducts. Epidemiological studies link drinking or bathing in chlorinated waters and bladder cancer,11 yet the identification of the chemical culprits has remained elusive and could be related to polypeptide-derived byproducts. Similarly, in the human bloodstream, neutrophils generate hypochlorous acid via myeloperoxidase enzymes as part of the immunological response.12 The reactions of hypochlorous acid can result in acute tissue damage or form proteinaceous byproducts that may be involved with inflammation-associated carcinogenesis. In all of these cases, determining the significance of amino acid side chain decomposition within polypeptides is hindered by the lack of an appropriate analytical technique for characterizing transformation products. While high resolution mass spectrometry techniques (e.g., MALDI-TOF MS) could theoretically characterize modifications to side chains within a discrete and well-characterized polypeptide,13 proteinaceous matter in natural waters and in the human body represents an assemblage of polypeptides differing in both the number and order of amino acid monomers.14 A technique involving polypeptide cleavage into the constituent amino acid monomers, followed by mass spectral analysis of the liberated monomers, would enable characterization of transformed amino acid residues. We developed a technique combining sample concentration by vacuum evaporation, polypeptide cleavage by enzymatic digestion, and characterization and quantification of modified amino acid monomers using high performance liquid chromatography mass spectrometry (HPLC-MS). Due to the low polypeptide concentrations in disinfected waters, freshwater, and seawater (i.e., < 1 mg/L), sample concentration is desirable to achieve lower detection limits. In the case of blood samples, large volumes of blood are generally not available, and polypeptide concentrations are sufficient (i.e., ∼10 g/L) that concentration may be unnecessary.15 Seeking to avoid any fractionation of polypeptides by hydrophobicity associated with concentration techniques such as extraction into hydrophobic solvents or resins, we evaluated sample concentration by vacuum evaporation using a lyophilizer. Although volatile byproducts are lost during lyophilization, polypeptides are relatively nonvolatile. While acid and base-catalyzed hydrolysis approaches have been used for polypeptide cleavage,16-18 the extreme pH and temperature conditions employed may promote analytical artifacts. For (11) Villanueva, C. M.; Cantor, K. P.; Grimalt, J. O.; Malats, N.; Silverman, D.; Tardon, A.; Garcia-Closas, R.; Serra, C.; Carrato, A.; Castano-Vinyals, G.; Marcos, R.; Rothman, N.; Real, F. X.; Dosemeci, M.; Kogevinas, M. Am. J. Epidemiol. 2007, 165, 148–156. (12) Witko-Sarsat, V.; Rieu, P.; Descamps-Latscha, B.; Lesavre, P.; HalbwachsMecarelli, L. Lab. Invest. 2000, 80 (5), 617–653. (13) Salavej, P.; Spalteholtz, H.; Arnhold, J. Free Radical Biol. Med. 2006, 40, 516–525. (14) Nunn, B. L.; Norbeck, A.; Keil, R. G. Mar. Chem. 2003, 83, 59–73. (15) Henry, R. J. Clinical Chemistry: Principles and Technics; Harper & Row: New York, 1964. (16) Cowie, G. L.; Hedges, J. I. Mar. Chem. 1992, 37, 223–238. (17) Cowie, G. L.; Hedges, J. I. Limnol. Oceanogr. 1992, 37 (4), 703–724. (18) Martens, D. A.; Loeffelmann, K. L. J. Agric. Food Chem. 2003, 51, 6521– 6529.

Figure 1. Structure of GFL peptide. Stars represent amino acids featuring carbons and nitrogens universally mass-labeled with 13C and 15N.

example, previous studies indicated that disinfectant reactions with model amines transform amines to nitriles19 and N-nitrosamines.20 However, low pH conditions may promote nitrosation in the presence of nitrite,21 while both low and high pH conditions may hydrolyze nitriles. We evaluated enzymatic digestion for peptide cleavage because the circumneutral pH conditions employed would minimize these artifacts. Pronase E, an enzyme cocktail derived from Streptomyces griseus digestion, is frequently used for proteolysis because it is hydrolytically active on many of the possible amino acid linkages.22,23 However, the influence of salinity, pH, and natural organic matter (NOM) loading on its digestion efficiency has been incompletely characterized. Using a multivariate statistical approach, we evaluated the efficacy of Pronase E liberation of amino acids from polypeptides across the range of pH, salinity, and NOM concentration conditions relevant to seawater, freshwater, wastewaters, disinfected drinking/ recreational waters, and human blood. The method was applied to characterize and quantify lysine nitrile, one of the predicted transformation products of the primary amine moiety within the lysine side chain of polypeptides in chlorinated drinking/ recreational waters. EXPERIMENTAL SECTION Materials. Poly-L-lysine, Pronase E (Protease type XIV), glycylglycyl-phenylalanine (GGF), glycyl-phenylalanine (GF), phenylalanyl-leucine amide hydrobromide (FL), and glycyl-glycine hydrochloride (GG), were obtained from Sigma-Aldrich. Phenylalanine (F) was obtained from Advanced ChemTech (Louisville, KY). GFL peptide, an octapeptide with a central glycyl-phenyl-leucine (GFL) tripeptide containing carbons and nitrogens universally masslabeled with 13C and 15N (Figure 1), was obtained together with its constituent universally labeled L-leucine from Cambridge Isotope Laboratories. Bovine serum albumin (BSA) served as a (19) Joo, S.-H.; Mitch, W. A. Environ. Sci. Technol. 2007, 41 (4), 1288–1296. (20) Schreiber, M. I.; Mitch, W. A. Environ. Sci. Technol. 2006, 40 (19), 6007– 6014. (21) Mitch, W. A.; Sharp, J. O.; Trussell, R. R.; Valentine, R. L.; Alvarez-Cohen, L.; Sedlak, D. L. Environ. Eng. Sci. 2003, 20 (5), 389–404. (22) Sweeney, P. J.; Walker, J. M. In Methods in Molecular Biology, Vol 16 Enzymes of Molecular Biology; Burrell, M. M., Ed.; Humana Press, Inc.: Totowa, NJ, 1993. (23) Brisbane, P. G.; Amato, M.; Ladd, J. N. Soil Biol. Biochem. 1972, 4, 51– 61.

Analytical Chemistry, Vol. 81, No. 18, September 15, 2009

7651

model protein in these studies and was obtained from American Bioanalytical (Natick, MA). The NOM surrogate was a humic acid fraction isolated from an Amherst (MA) peat soil as described elsewhere.24 Free chlorine and monochloramine stock solutions were generated and standardized as described previously.25 Synthesis of Lysine Nitrile (2-Amino-5-cyanopentanoic acid). The use of alcoholic hypochlorite, as described in Yamazaki for primary amines,26 for the cyanation of the ε-amine of lysine requires base-stable protection of the R-amine. Accordingly, 50 mg (0.2 mmol) of boc-protected lysine (Chem-Impex, Wood Dale, IL), (S)-6-amino-2-(tert-butoxycarbonylamino)hexanoic acid, was dissolved in 15 mL of methanol and 50 µL of tert-butanol, which was found to enhance dehydrohalogenation. Calcium hypochlorite (100 mg, 1 mmol) was then added, and the mixture was stirred rapidly at 40 °C for 24 h. Water and pentane (8 mL each) were then added sequentially, and the pH was neutralized with 0.1 M hydrochloric acid. A vortex mixer was used to extract the nonpolar byproducts (e.g., mono- and dihalogenated amine intermediates) into pentane as waste. The aqueous phase was then concentrated to 1 mL with a Savant SVC100H speed vac, adjusted to pH 4.5 with 0.1 M hydrochloric acid, and loaded onto a 12 mL Supelco DSC-18 solid phase extraction (SPE) cartridge that was preconditioned as below. Cartridges were then rinsed with water (3 × 5 mL) to remove salt. Unreacted starting material and boc-lysine nitrile, 2-(tert-butoxycarbonylamino)-5-cyanopentanoic acid, were eluted with three 8 mL portions of 0.05% formic acid in 20% acetonitrile that were collected, combined, and concentrated to dryness. Semipreparative HPLC (Shimadzu LC-10AT) with UV detection at 205 nm (Shimadzu SPD-10A) was used for subsequent purification of boc-lysine (Rt 8.2 ± 0.5 min) and boc-lysine nitrile, (Rt ) 11.2 ± 0.3 min); the mobile phase composition was isocratic 5% acetonitrile in 10 mM ammonium formate with a 4.5 mL/min flow rate and an Econosphere C18-10 µ column (Alltech 28085). Fractions containing boc-lysine nitrile were pooled, concentrated, and desalted as described above with 20% acetonitrile eluent. After a final concentration from ∼8 mL to 200 µL, trifluoroacetic acid (60 µL) was added as a boc deprotection agent. Lysine nitrile was recovered in ∼85% overall yield following 30 h in vacuum at 40 °C: colorless; 1H NMR (500 MHz, DMSOd6, δ) 1.55-1.8 (4H, m, H2-3, H2-4), 1.85-1.92 (2H, m, H2-5), 3.95 (1H, t, 3J ) 5.3, 5.9 Hz, H-2). HRESIMS (m/z): [M - H]- calcd for C6H9 N2O2, 141.0742; found, 140.9549. Routine Sample Processing Procedure. Water samples were collected in fluorinated high-density polyethylene containers, and the pH and temperature of each was recorded. Total residual chlorine was measured by the DPD colorimetric method,27 and the dissolved organic carbon was measured with a Shimadzu TOC analyzer. After quenching residual chlorine by addition of 200 mg of ascorbic acid, samples were refrigerated pending analysis. (24) Pignatello, J. J.; Kwon, S.; Lu, Y. Environ. Sci. Technol. 2006, 40, 7757– 7763. (25) Schreiber, I. M.; Mitch, W. A. Environ. Sci. Technol. 2005, 39 (10), 3811– 3818. (26) Yamazaki, S. Synth. Commun. 1997, 27 (2), 3559–3564. (27) Eaton, A. D., Clesceri, L. S., Greenberg, A. E., Eds. Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association, American Water Works Association, Environment Federation Publishers: Washington D.C., 1998.

7652

Analytical Chemistry, Vol. 81, No. 18, September 15, 2009

Samples were supplemented with 100 µL of a 10 g/L solution (1 mg) of GGF, the commercially inexpensive central tripeptide within GFL (Figure 1), to correct for losses during concentration and subsequent digestion. The samples were evaporated directly within glass lyophilization jars to ∼8 mL, and then, the sample was transferred to 40 mL amber glass vials for enzymatic digestion as described below. The ∼8 mL concentrate was supplemented with calcium chloride (1 mL of a 22 g/L stock solution) to stabilize Pronase E.22 The pH was adjusted to 8.1 (±0.1) with sodium hydroxide or hydrochloric acid. Pronase E stock solutions (10 g/L) were made fresh daily in deionized water, and 1 mL of this stock (10 mg) was added to each concentrate. After adding deionized water to bring the final volume of the concentrate to 10 mL, 10 µL of a 1 mM GFL peptide stock solution (10 µg) was added to evaluate digestion efficiency. The solutions were capped, mixed, and incubated at 37 °C for 30 h. After incubation, the vials were cooled to 25 °C and the digested concentrate was transferred to 12 mL Supelco Discovery C18 DSC-18 SPE cartridges containing 2 g of extraction material. The cartridges had been preconditioned with sequential acetonitrile (2 × 10 mL), 50%:50% acetonitrile/water (2 × 10 mL), and water (4 × 10 mL) rinses. After loading with the concentrate, the cartridges were eluted with a series of water/ acetonitrile mixtures (20 mL each) into separate 50 mL Falcon tubes: 100:0, 80:20, 60:40, 40:60, 20:80, 10:90, 10:90. Eluted fractions were concentrated to dryness via vacuum evaporation and reconstituted in 1 mL of 70:30 water/acetonitrile. The fractions were fortified with 5 µL of a 1000 ng/µL boc-lysine in isopropanol stock solution as an external standard prior to HPLC-MS analysis. Chemical verification was based on chromatographic retention times and mass spectrometry. Analytes were quantified on the basis of single ion monitoring and/or extraction for each fraction, and concentrations were reported herein as sum totals, unless otherwise noted. Complementary verification of lysine nitrile was with MS2 or high-resolution techniques. HPLC-MS Analyses. A low resolution MS (System 1) was employed for all analyses with the exception of quantification of lysine nitrile in chlorinated drinking water and pool water samples; for the latter samples, a high resolution MS system was employed (System 2). System 1. A Surveyor MS pump in tandem with a Finnigan LCQ Deca ion trap mass spectrometer (HPLC-MS), fitted with a Varian Inertsil ODS-3 analytical column (L ) 250 mm, ID ) 4.6 mm, S ) 5 µm) and a ThermoFinnigan Surveyor Plus photodiode array detector (PDA), were used for the evaluation of analyte recovery and proteolytic efficiency. Full-loop injections (20 µL) were made with a column flow of 1 mL/min composed of three eluents: (a) methanol, (b) 10 mM ammonium formate in deionized water, and (c) 10 mM ammonium formate in a 10%:90% deionized water/acetonitrile mixture. The mobile phase composition (in terms of percent ratios) was isocratic at 4a:91b:5c for 10 min, ramped to 4a:0b:96c over 11 min, held for 2 min, ramped back down to 4a:91b:5c over 7 min, and held for 8 min. Column effluent was routed through the PDA (205-400 nm) and to the mass spectrometer. Spectra were obtained using electrospray ionization in the positive mode with a (±)4.5 kV spray voltage, a 4.5 mm i.d.

Table 1. Four Factors and Five Factor Levels Used in the Central Composite Multivariate Experimental Design factor levels factor (units)

-R

-1

cpb

1

R

x1: salt load (g/L) NH4HCO3:CaCl2 mM ratio x2: pronase E to proteina ratio (weight %) x3: humic NOM (mg/L) x4: pH

0 0.1

1.2 10:4 0.5

6.2 50:20 1.0

31.1 250:100 5.0

62.0 500:200 10.0

0 7.0

250 7.5

500 8.0

750 8.5

1000 9.0

b

a Total protein ) 99 mg of BSA, 1 mg of GGF, and 10 µg of GFL. cp ) center point.

capillary sample tube, and a 275 °C capillary temperature. Nitrogen sheath and sweep gas flow rates were 40 and 20 arb, respectively. Three select ion monitoring (SIM) ionization segments were used in series for quantitation: m/z (+) 131.00-145.00, 175.50-178.50 (0-5.75min);m/z(+)160.00-180.00,220.00-240.00,275.00-285.00 (5.75-12.02 min); and m/z (+) 242.00-249.00, 294.50-297.50, 299.50-300.50, 1000.50-1003.50 (12.02-35.00 min). Table S1 in the Supporting Information provides the retention times and quantification ions for GGF, GFL, their digestion products, and lysine nitrile. System 2. Dual Shimadzu LC-10AD pumps, a Shimadzu SC10A PDA, a YMC-Pack ODS-AQ analytical column (L 250 mm, i.d. 4.6 mm, S 5 µm), and an Applied Biosystems QSTAR XL high resolution quadrupole time-of-flight (QTOF) mass spectrometer were used. The mobile phase composition, flow rate, and relative retention of analytes were as above. After the PDA, however, a splitter directed ∼90% of the flow to a collection reservoir and ∼10% (100 µL/min) to the MS. Spectra were obtained over m/z (±) 80-800 using an ion spray source with spray voltages of +4750 or -4500, declustering potentials of ±70, ion source and curtain gas flow rates (arb) of 1.13 L/min, an ion release delay of 6 (arb), and an ion release width of 5 (arb). Focusing potentials were -280 and +230 for negative and positive ionization, respectively. Proteolytic Efficiency: Experimental Design. We evaluated the influence on polypeptide digestion efficiency by Pronase E in model water concentrates as a function of four experimental factors: salinity (x1), pH (x2), NOM (x3), and Pronase E (x4). Within a final volume of 10 mL of deionized water, all model concentrates contained 1 mL of a 99 mg/mL BSA solution (99 mg) as a model protein, 100 µL of a 10 mg/mL GGF peptide solution (1 mg), and 10 µL of a 1 mM GFL peptide solution (10 µg) to correct for digestion efficiency. The ∼100 mg of polypeptide approximates the protein content in 10 mL of blood.15 However, it far exceeds the protein content of drinking waters, where the protein content of 1 L samples approaches 100 µg.28 A multivariate experimental array, designed with Design-Expert 7.0 (Stat-Ease Inc., Minneapolis, MN), was used to evaluate the influence of the four experimental factors (x1-x4). The experimental array featured a central composite design with different combinations of the four experimental factors, each adjusted over five levels (Table 1). A total of 30 experiments was run simultaneously, with six replicates for the center-point conditions. (28) Dotson, A.; Westerhoff, P. J. AWWA 2009, in press.

Salinity was varied between 0 and 62 g/L using combinations of ammonium bicarbonate (NH4HCO3) and calcium chloride (CaCl2) maintained at a constant 5:2 molar ratio of NH4HCO3: CaCl2. These salts are used in a previous study of Pronase E stability,22 and calcium and bicarbonate are important constituents of natural waters. This salinity range brackets the levels expected for unconcentrated blood and seawater or a 100-fold concentration of freshwaters. NOM concentrations were varied from 0 to 1000 mg/L (∼0-400 mg/L as dissolved organic carbon (DOC)) using the Amherst humic acid isolate, capturing the upper limit expected for a natural water concentrated 100fold. Solution pH was varied between 7.0 and 9.0. Pronase E ranged from 0.1 to 10 mg, representing Pronase E to protein weight ratios of 0.1-10%. The center-point conditions, with the exception of NOM loading, were based on concentrations/ loadings previously found to produce the maximum Pronase E digestion efficacy.22 Reagents were added in the following order to promote solubilization in the presence of the elevated ammonium bicarbonate concentrations: protein, NOM, calcium chloride, and ammonium bicarbonate. The pH was adjusted using 0.1 M hydrochloric acid or sodium hydroxide. Finally, an appropriate volume of freshly prepared Pronase E solution (10 mg/mL) was added. Vials were incubated for 30 h at 37 °C with gentle mixing and analyzed as described above. RESULTS AND DISCUSSION Sample Concentration and Matrix Recovery Quality Control. Seeking to avoid any fractionation of polypeptides by hydrophobicity associated with concentration techniques such as extraction into hydrophobic solvents or resins, we evaluated sample concentration by vacuum evaporation using a lyophilizer, during which only volatile byproducts are lost. However, loss of polypeptides to the lyophilization vessel wall was a concern. To evaluate these losses, 3 L samples were prepared and split into 1 L subsamples that yielded the multivariate experimental centerpoint concentrations after evaporative concentration of each to 10 mL. GGF, the commercially inexpensive central tripeptide within GFL (Figure 1), served as a useful diagnostic peptide, and we added 100 µL of a 10 g/L solution (1 mg) to two of the subsamples, one prior to concentration and one after concentration. A comparison of GGF recovery on the undigested subsamples was then made using HPLC-MS. In general, GGF recoveries used to diagnose peptide losses to the lyophilization vessel walls indicated