Normal human urine is generally believed to be almost free of protein and polypeptides1,2,3. This is thought to be due to selective glomerular permeability and efficient tubular uptake of protein filtered by the glomerulus1. There is much molecular, physiologic, and anatomic evidence that glomerular ultrafiltration is highly selective. Although exact measurements are not available, micropuncture studies in the rat suggest an albumin concentration in the glomerular filtrate of 22.9 mg/L4, and similar estimates have been made in the dog5. An indirect estimate in man, based on studies of the renal Fanconi syndrome, gave a value of approximately 3.5 
0.41 mg/L6. Using this estimate, human plasma, which contains some 40 g/L of albumin, forms an ultrafiltrate containing approximately 600 mg albumin over a 24-hour period. Over 95% of albumin is then removed from this ultrafiltrate, probably by megalin/cubilin-dependent endocytosis in the proximal tubule, to produce urine with an albumin content of less than 30 mg/24 hr7,8,9. When this endocytic uptake is defective, as in the Fanconi syndrome, albumin and other proteins in the ultrafiltrate are found in urine and constitute tubular proteinuria10,11.
In the Fanconi syndrome (Lignac-de Toni-Debré-Fanconi syndrome), reuptake of filtered proteins by the proximal tubule is impaired, and the resulting tubular proteinuria comprises about one half albumin, one third low-molecular-weight proteins (retinol-binding protein, 
2-microglobulin, and 
1-microglobulin), and the remainder other proteins10,11. The renal Fanconi syndrome is found in patients with Dent's disease (CLCN5 mutation), the oculocerebrorenal syndrome of Lowe (Lowe syndrome), as well as cystinosis and several other inherited and acquired diseases10,12,13.
The above description, based partly on protein measurements using a variety of chromogenic total protein assays and in-gel protein detection reagents, as well as immunochemical estimation of individual proteins, particularly albumin, has been challenged14. A large body of human and animal data from Comper et al14 has been interpreted to show that the normal glomerulus is relatively leaky, and passes 100 to 600 g of plasma protein per day into the glomerular filtrate. It is believed that a proportion of the albumin is partially degraded and then re-excreted into the tubular lumen to produce about 1 to 2 g of polypeptide fragments of mass 500 Da to 10 kD in the final urine. These fragments are believed to be undetectable by chromogenic protein reagents, such as Coomasie Blue-based dyes, as well as being unreactive in immunoassays for the parent protein, such as albumin14,15. Correct interpretation of the available data is critical for an understanding of both normal renal physiology and for hypotheses that implicate proteinuria itself, or peptiduria, in the progression of renal disease. This has been a controversial issue3,16,17.
In this study, we have defined proteins as polypeptides of mass >10 kD, and use the terms polypeptide and peptide interchangeably, with a definition of the mass-range where each term is used.
A critical part of the Comper hypothesis is the interpretation that chromogen formation by urine with the Biuret reagent is due to polypeptides only15. To provide experimental data that is more specific than the Biuret reaction, we have re-examined protein and polypeptide excretion in human urine in terms of its basic chemical constituents, that is, amino acid content, following size-exclusion chromatography (SEC) Figure 1. We have determined conditions where there is good overall recovery of amino acids so that no major urinary component will be omitted.
Figure 1.
Flow-chart showing the experimental design. Size exclusion chromatography (SEC) and amino acid analysis is shown on the left, and proteomic analysis on the right.
Full figure and legend (68K)To complement the technique of SEC and amino acid analysis we have used a proteomic mass spectrometric approach in which urinary proteins and polypeptides are isolated from urine and then analyzed by mass spectrometry (MS) Figure 118. The MS analysis is performed either on intact polypeptides using matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (MALDI-TOF MS), or by analysis of the tryptic digests of these polypeptides using nano-flow liquid chromatography directly coupled to electrospray ionization/tandem mass spectrometry (NanoLC-ESI-MS/MS)18. Although the mass spectrometric approaches do not have the quantitative power of amino acid analysis, they permit definitive qualitative identification of urinary polypeptides.
As well as normal urine, we have examined Fanconi urine. This acts as a control because protein excretion is known to be greatly elevated and also comprises a group of patients in which previous immunochemical studies have shown increased excretion of several potentially bioactive peptides, which might contribute to the progression of renal failure6,10,18
The complementary approaches of amino acid analysis and mass spectrometry have produced data that are mutually consistent, and do not support the hypothesis that there are large quantities of polypeptide fragments excreted in normal human urine. Furthermore, the data suggest that even in Fanconi urine, intact proteins predominate, although there is a large increase in excretion of polypeptides in the mass range 2 to 5 kD.
METHODS
The flow chart shown in Figure 1 summarizes the methods used for SEC with amino acid analysis and for the mass spectrometric experiments.
Patients and controls
Urine samples were collected as described previously11, and stored in liquid nitrogen until analysis. The patients with Fanconi syndrome consisted of: 2 with Dent's disease, patients C/II/2 with a total CLCN5 deletion, and F/II/1, W279X16,19, and 3 with Lowe syndrome. The 3 patients with Lowe's syndrome had severe mental and growth retardation, visual impairment, and tubular proteinuria. The calculated creatinine clearance for the 5 patients with Fanconi syndromes was >60 mL/min (4 patients), except for 2 Lowe syndrome patients, 33 and 37 mL/min (data not shown)20. The control patients (4 male and 1 female) had no personal or family history of renal disease, and albumin content of their urine measured immunochemically was normal (data not shown).
Preparation of urine samples
An aliquot of urine was taken for creatinine estimation (Dade-Behring Dimension® alkaline rate picrate method, CREA DF33A; Dade-Behring, Deerfield, IL, USA) and aliquots were centrifuged at 200,000g for 30 minutes at 4°C to remove Tamm-Horsfall glycoprotein.
Size-exclusion chromatography
After ultracentrifugation, 5 mL supernatant was applied to a 40 
2.5-cm column of Superdex® 30 prep grade agarose-dextran medium (catalog no. 17-0905-10, Amersham Biosciences, UK, Ltd., Buckinghamshire, UK) at 3 to 6°C, equilibrated with 0.22 
-filtered 0.1 mol/L ammonium bicarbonate (Fluka Brand, catalog no. 09830; Sigma-Aldrich Co., Ltd., Dorset, UK) at a flow rate of 40 mL/h. Fractions of 4 mL were collected and frozen at -80°C. The column was cleaned with 0.5 mol/L sodium hydroxide between each analysis.
Fractions that corresponded to the molecular mass ranges shown in Figure 1 were pooled and lyophilized to remove the volatile chromatography buffer. Finally, each pool was redissolved in 1 to 2 mL water, split into 2 equal portions, and relyophilized. For amino acid analysis, 1 of these 2 samples was redissolved in 0.5 to 1 mL water.
Superdex® 30 columns were calibrated with 6 molecular weight markers (catalog nos. are Sigma-Aldrich unless stated): 2-microglobulin, MW 11.6 kD (SCIPAC, Ltd., Kent, UK); aprotinin, catalog no. A4529, 6.5 kD; insulin B chain, catalog no. I6383, 3.5 kD; gastrin 1, catalog no. G9020, 2.1 kD; substance P, catalog no. S6883, 1.62 kD, and Asp-Asp-Asp-Asp, catalog no. A4440, 478 Da.
In preliminary experiments, columns of either Toyopearl® HW-40F (Sigma-Aldrich Co., Ltd.) 40 2.5 cm, or Sephadex® G-10 (Amersham Biosciences, UK, Ltd.) 80 1.6 cm, substituted for the Superdex® 30 column, but other conditions were the same.
Trypsin digestion of albumin
Human albumin, Sigma A8763, 2 mg in 2 mL 0.1 mol/L ammonium bicarbonate, was incubated with 100-g bovine trypsin, Sigma HPLC grade, T8658, for 36 hours at 37°C, and the reaction terminated by addition of excess 4-(2-aminoethyl)benzenesulfonyl fluoride, Calbiochem catalog no. 101500. The solution was applied to a Superdex® 30 column as above, and amino acid analysis performed on peptide fractions in the same way. In preliminary experiments measuring immunoreactive albumin, these conditions gave complete digestion by the trypsin.
Amino acid analysis after acid hydrolysis
An appropriate volume of each sample and of L-norleucine internal standard was transferred to a 4-cm 3-mm microtube, previously cleaned by heating to 500°C for 15 hours, and concentrated to dryness. Gas phase acid (hydrochloric) hydrolysis was performed at 115°C for 22 hours. After removing traces of acid, the residue was dissolved in sodium citrate loading buffer (pH 2.2), and filtered by centrifugation through a 0.2- filter. An aliquot of the filtrate was injected into a loading capsule placed in a Pharmacia Alpha Plus® series amino acid analyzer (Biochrom Ltd., Cambridge, UK), and chromatography performed on a sodium system ion exchange resin eluting with buffers over the pH range 3.2 to 6.45. Peak detection was achieved by mixing the eluate with ninhydrin at 135°C and measuring the absorbance at 570 and 440 nm.
Physiologic amino acid analysis without acid hydrolysis
L-norleucine internal standard was added to the pooled fractions corresponding to free amino acids and lyophilized. This was redissolved in 100 L water, 100 L 10% sulphosalicylic acid was added, and the solution incubated for 60 minutes at 4°C and centrifuged at 3000g for 15 minutes. The supernatant, 60 L, was mixed with 45 L 0.3 mol/L lithium hydroxide, and 17.5 L was loaded into a capsule on a Biochrom 20® amino acid analyzer (Biochrom Ltd.).
SDS gel electrophoresis and protein staining
The pellet derived by ultracentrifugation of 1.5 mL urine was solubilized in Laemmli sample buffer as described previously21, except that 100 mmol/L dithiothreitol was added to sample buffer. Samples were heated for 4 minutes in a boiling water bath, and 10% to 20% Tris-HCl gels (Bio-Rad, Hemel Hempstead, UK) were used to separate proteins that were detected by both silver and Coomasie Blue staining (Sigma AG-5 and B8522 used according to supplier's instructions).
Extraction of peptides from urine by reversed phase (RP) and strong cation exchange (SCX) chromatography
Urine was centrifuged at 3000g for 15 minutes and the supernatant mixed with an equal volume of 5% acetonitrile (ACN)/0.1% trifluoroacetic acid (TFA), and injected onto a C18 column (50 0.8 mm) packed with POROS R2 20 RP (Applied Biosystems, Warrington, UK). The column was washed with 20-column vol. 4% ACN/0.1% TFA, and peptides eluted directly onto a POROS HS SCX column (50 0.8 mm) (Applied Biosystems) using 150 L of 80% ACN/0.1% TFA. Bound peptides were washed with 20-column vol. 4% ACN/0.1% TFA, eluted with 150 L of 2 mol/L ammonium acetate in 25% ACN/0.1% TFA, dried in a Speed Vac, redissolved in 5% ACN/0.1% TFA, and protein quantitated by Bradford assay.
Reversed-phase HPLC (RP-HPLC) of urine proteins and peptides
Solvent A was 0.1% TFA, and solvent B 80% ACN/0.1% TFA. Polypeptides (5 g) from SCX were injected onto a C18 POROS R2 10 column (250 mm 320 m) eluted at 20 L/min, and washed with the following% solvent B with the balance solvent A: 5% B for 8 minutes, 15% B for 1 minute, 15% to 60% B in 27 minutes, and finally 100% B for 20 minutes.
MALDI-TOF MS of RP-HPLC fractions
RP-HPLC fractions were dried, redissolved in 10 L water, and 0.5 L mixed with 1 L of 2,5-dihydroxybenzoic acid (DHB) on a MALDI plate, dried, and MALDI-TOF mass spectra–collected in an Ultraflex TOF/TOF mass spectrometer (Bruker Daltonik, Bremen, Germany).
NanoLC-ESI-MS/MS
Five microliters of each RP-HPLC fraction redissolved as above was added to 10 L of 50 mmol/L ammonium bicarbonate containing 100 ng of sequencing grade trypsin (Promega, Madison, WI, USA) and incubated overnight at 37°C. NanoLC-ESI-MS/MS of the resultant tryptic peptides was carried out with an Ultimate HPLC (LC Packings, Amsterdam, The Netherlands) connected on-line to a Q-TOF mass spectrometer (Micromass, Manchester, UK).
MALDI-TOF MS of SEC fractions of 
750 Da
SEC fractions corresponding to molecular masses 750 Da were pooled, lyophilized, desalted using an SPE cartridge (Oasis; Millipore, Billerica, MA, USA), and loaded on an SCX column equilibrated with 0.5% acetic acid/20% CAN, and the absorbance at 280 nm allowed to return to baseline. The eluate was collected and referred to as the acidic fraction. The mobile phase was then changed to 500 mmol/L ammonium acetate dissolved in 0.5% acetic acid/20% CAN, and the eluate referred to as the basic fraction. The acidic and basic fractions were analyzed by MALDI-TOF MS and NanoLC-ESI-MS/MS, as above.
Nonpeptide Biuret chromogens
A mixture of the following was made in 50 mmol/L sodium phosphate pH 7.00: urea, 330 mmol/L (Sigma U0631); creatinine 13.3 mmol/L (Sigma C4380); ammonium chloride, 36 mmol/L (Merck 10017); uric acid, 3.1 mmol/L (Sigma U2875); D-glucose, 0.4 mmol/L (Sigma G7021); L-histidine, 1 mmol/L (Sigma H8125); L-cystine, 0.1 mmol/L (Sigma C8755); L-asparagine, 0.2 mmol/L (Sigma A0884); L-serine, 0.45 mmol/L (Sigma S4500); L-threonine, 0.2 mmol/L (Sigma T8625); urobilin, 1 mg/L (Porphyrin Products U590–9); hippuric acid, 9 mmol/L (Sigma H6529); 
-methyl-L-histidine, 0.34 mmol/L (Bachem F2480); and phenyl glucuronide, 2 mmol/L (Fluka 78555).
RESULTS
Table 1 and Figure 2 show the total amino acid content of normal and Fanconi urine in the mass range from >10 kD (protein) to <750 Da (free amino acids or small peptides). Amino acid content refers to amino acid content after acid hydrolysis as described in Methods, unless stated otherwise. The complete chromatographic profile of normal urine is shown in Figure 3a. To examine amino acid content in more detail, these data are also shown in Figure 3b, rescaled by a factor of 80 and omitting mass <250 Da, which comprise mostly free amino acids (see below). Figure 3b confirms the relative absence of intermediate mass material. Quantities of each amino acid in normal urine are given in Table 2.
Figure 2.
Comparison of mean total amino acid excretion (after acid hydrolysis) between normal (
) and Fanconi (
) urine, mg per 24 hr in 5 mass ranges,N = 5for each group. Bars show SEM.
Figure 3.
(A) Complete amino acid content (after hydrolysis) of mass fractions isolated from normal urine by size-exclusion chromatography. Results are means from the analysis of urine specimens from 5 individuals. Amino acids are expressed as mol/mmol creatinine. (B) As for (A), but omitting the fraction of <250 Da, which contains primarily free amino acids, and with 80-fold increase in scale. Abbreviations for amino acids are: G, glycine; D+N, L-aspartic acid and L-asparagine; T, L-threonine; S, L-serine; E+Q, L-glutamic acid and L-glutamine; A, L-alanine; V, L-valine; M, L-methionine; I, L-isoleucine; L, L-leucine; Y, L-tyrosine; F, L-phenylalanine; H, L-histidine; K, L-lysine; R, L-arginine; HO-K, hydroxylysine; MeHis, 1- and 3-methyl-L-histidines; HO-P, hydroxy-L-proline; P, L-proline.
Table 1 - Total amino acid content (after hydrolysis) of mass fractions isolated from normal and Fanconi urine by size-exclusion chromatography.
Full table Table 2 - Amino acid content (after hydrolysis)a of mass fractions isolated from normal and Fanconi urine by size-exclusion chromatography.
Full table As anticipated, Fanconi urine demonstrated gross elevation of the amino acid content corresponding to mass >10 kD (Tables 1 and 2, Figures 2 and 4). In addition, large increases in the excretion of smaller peptides compared to normal urine were also found (Table 2 and Figure 4). Although the molecular mass ranges stated in Tables 1 and 2 and Figures 2, 3, 4 are approximate, negligible cross contamination of the mass range of 2 to 5 kD with material of >10 kD or of <750 Da was found in control experiments using peptide mass standards (data not shown). This suggests that in the Fanconi syndrome there is an approximately 12.9-fold elevation of peptides in the 2 to 5 kD range, compared with normal urine (Table 1 and Figure 4).
Figure 4.
Complete amino acid content (after hydrolysis), 
mol/mmol creatinine, of mass fractions isolated from Fanconi urine by size-exclusion chromatography. Results are the mean from the analysis of urine specimens from 5 patients. Amino acids are expressed as mol/mmol creatinine. Abbreviations for amino acids are as in Figure 3.
Ultracentrifugation of urine before column chromatography was required to eliminate interference from Tamm-Horsfall protein (uromodulin) and particulates. Unless Tamm-Horsfall protein is removed from the urine, it forms a gel barrier on the surface of the column, impairing the chromatography. Ultracentrifugation caused negligible losses of total amino acids measured after hydrolysis, recovery (mean SEM) in the supernatant following ultracentrifugation being 103.1 5.3% and 97.5 5.8% of the original normal and Fanconi urine, respectively (N = 5 for each). Examination of the pellet proteins from this step by reducing SDS gel electrophoresis and staining with either Coomasie dye or silver showed that the pellet contained a virtually pure protein that comigrated with authentic Tamm-Horsfall protein (data not shown).
Initial experiments using the 3 size-exclusion media as described under Methods established that only the agarose-dextran column gave satisfactory overall recovery and resolution of amino acids from urine; only this medium was used subsequently. Overall recovery of amino acids from urine was 73.4 4.1% and 76.2 12.4% for normal and Fanconi urine, respectively (N = 5 for each, defined as mean SEM total amino acid content after hydrolysis in all SEC fractions compared to the quantity of amino acids applied to the column). No correction has been applied to losses due to hydrolysis (other than any loss to the L-norleucine internal standard following its addition to post-chromatography fractions), or to sample handling.
To provide further information about recovery following SEC, the following pure proteins, of molecular weight shown, were each applied to the agarose-dextran column used for urine studies, and the overall recovery of protein determined and given as a percentage of that applied: 2-microglobulin, MW 11.6 kD, 90.7%; aprotinin, 6.5 kD, 69.9%; insulin B chain, 3.5 kD, 80.6%; insulin A chain, 2.5 kD, 63.6%; gastrin 1, 2.1 kD, 75.2%; substance P, 1.62 kD, 90.9%. Additionally, we exhaustively digested human serum albumin with trypsin, and determined the recovery of total amino acids following acid hydrolysis using the same column. This gave a peptide elution profile ranging from >10 kD to <750 Da (data not shown); overall recovery of amino acids was 94%, 72% of which comprised peptides of <10 kD.
Physiologic amino acid excretion, that is, amino acid excretion without acid hydrolysis, was measured in SEC fractions corresponding to free amino acids. (As described in Methods, a different amino acid analysis protocol was required from that used for the amino acid analyses after acid hydrolysis described above.) Results of physiologic amino acid excretion in normal and Fanconi urine are shown in Figure 5. In contrast to the approximately 12.9-fold elevation of peptides in the 2 to 5 kD range in Fanconi compared to normal urine, free amino acid excretion was only 2.14 0.73-fold elevated (N = 5 for each).
Figure 5.
Comparison of mean physiologic amino acid excretion (no acid hydrolysis), mol/mmol creatinine, between normal () and Fanconi urine (),N = 5for each group. Bars show SEM.
NanoLC-ESI-MS/MS of basic and acidic fractions of the SEC fractions of molecular weight 750 Da did not identify any plasma protein–derived peptides, although peptide sequences derived from Tamm-Horsfall protein, -actin, and aminopeptidase N, all renal proteins, were detected in the basic fractions (data not shown).
To complement the approach based on size-exclusion chromatography and amino acid analysis, we quantitatively extracted polypeptides from urine by a combination of reverse phase (RP) and Strong Cation Exchange (SCX) chromatography. Figure 6 shows the absorbance profile at 214 nm when equal amounts of these polypeptides, isolated from normal and Fanconi urine, were subjected to RP chromatography with ultraviolet monitoring and gradient elution. The isolated fractions were examined by both MALDI-TOF MS for intact molecular weight determination, and NanoLC-ESI-MS/MS after tryptic digestion for amino acid sequence determination.
Figure 6.
Chromatograms of normal and Dent's urinary polypeptides eluted from a C18-POROS10 column. Protein (5 g) from the urine of (A) normal individuals () and (B) Dent's patients () was separated by capillary HPLC as described in Methods. Peaks marked with an asterisk have the same retention time as bovine serum albumin. In (C), the area under each fraction was calculated and plotted normalized to total area. Error bars represent SEM (t test, P > 0.05, N = 3).
MALDI-TOF-MS of SCX fraction 8 revealed the presence of relatively large polypeptides, including 2 m/z values that corresponded to albumin (m/z 67,800 singly, and 33,900 doubly charged), and had the same retention time as bovine serum albumin Figure 7. We sequenced 33% of albumin by NanoLC-ESI-MS/MS in fraction 8 of normal urine Figure 7c, and peptides from the N-, as well as C-termini, were detected. These data suggest that the major form of albumin present in both Dent's and normal urine is the full-length protein.
Figure 7.
Polypeptide patterns from the major RP-HPLC fraction in Figure 5. High molecular weight MALDI-TOF-MS spectra of normal (A) and Dent's (B) fraction 8. (C) Sequence of human serum albumin (NCBI gene identifier 3212625). Underlined peptides were automatically sequenced by NanoLC-ESI-MS/MS, and correspond to 33% of the total sequence.
Full figure and legend (75K)Previous work by Comper et al15 suggests that use of the Biuret reagent as a potential peptide-bond-specific detection reagent reveals high peptide levels in normal urine. We investigated the specificity of this reagent under the same conditions previously reported for its use to detect protein fragments, particularly using the same ratio of test fluid (urine) to Biuret reagent15. To do this, we first screened over 50 nonpeptide compounds known to be present in normal human urine22 at approximately 10-fold physiologic concentrations, and then selected 14 compounds from these (as stated under Methods), which gave significant absorbance at the wavelength of the Biuret chromogen. A mixture of these compounds at physiologic concentrations demonstrated an apparent albumin-equivalent concentration of approximately 0.4 g/L. This result demonstrates a lack of specificity of the Biuret reagent under the conditions used, and suggests that a significant proportion of apparent urinary peptide is due to nonpeptide chromogens.
To further examine previous reports that significant quantities of radiolabeled peptides could be separated from the urine of humans given human serum albumin tritiated in L-lysine, we cochromatographed radiolabeled L-lysine with different peptides, using conditions on Sephadex G-100® gel filtration medium very similar to those reported previously23. Although, for example, aprotinin (MW 6.5 kD) was eluted before free L-lysine (MW 146 Da), an octapeptide from cholecystokinin (Sigma catalog no. C2901; fragment 26–33, MW 1.06 kD) was eluted just after free L-lysine (results not shown). This result suggests that chromatography on Sephadex G-100® (protein size-exclusion limit 150 kD) may give poor resolution in separating small peptides from free amino acids.
DISCUSSION
If it could be shown that normal human urine contains gram quantities of plasma protein-derived fragments, much previous data on glomerular filtration and tubular protein handling would need to be reinterpreted1,3,9; indeed, this has been proposed14.
We have used 2 independent experimental approaches, size-exclusion chromatography with quantitative amino acid analysis, and 2 mass spectrometric approaches Figure 1. Both techniques have failed to detect the gram quantities of protein fragments in normal urine predicted by the Comper hypothesis14. Normal human urine, excluding Tamm-Horsfall protein, was found to contain 33.7 10.7 mg protein per 24hr (mean SEM), protein defined as polypeptide >10 kD, and peptide content in the range 750 Da to 10 kD was 22.0 9.6 mg. We believe this is the first report of these important nephrologic measurements in absolute amounts. Fanconi patients excreted greatly increased amounts of both protein 1740 660 mg/24hr and peptide 446 145 mg/24hr. Although incomplete resolution of the 5 to 10 kD peptide fraction from certain proteins is to be expected, peptides 2 to 5 kD were negligibly contaminated in this way and were elevated in 12.9 3.9-fold excess in Fanconi compared with normal urine. In contrast, free amino acid excretion was elevated by only by 2.14 0.73-fold in the 5 patients studied with Fanconi syndromes.
Following the initial ultracentrifugation step, recovery of total amino acid content measured after hydrolysis was quantitative (Figure 1 and Results); this is despite the loss of Tamm-Horsfall protein. This good recovery is probably explained by this protein comprising some 3% of total urinary amino acids (24 and Table 1) and the removal of particulates.
We selected conditions for SEC and amino acid analysis that gave good recovery of amino acids. Overall, 70% recovery from urine was achieved without any correction for losses, except hydrolytic losses of the L-norleucine internal standard. Purified peptides and polypeptides over the mass range 1.62 to 11.6 kD were recovered in yields from over 60% to 90%, and tryptic peptides from human serum albumin were recovered in an overall yield of 94%. Peptides of 750 Da or smaller were resolved from larger peptides on the SEC system employed. Thus, it is highly unlikely that such plasma-derived protein fragments in gram quantity could remain undetected by this approach.
The mass spectrometric approaches also failed to identify peptides in normal human urine that might be derived from plasma proteins. Those peptides that were identified in normal urine were probably of renal origin and had sequences corresponding to Tamm-Horsfall protein, -actin, and aminopeptidase N. Most albumin peptides detected probably originated from the full-length protein, a conclusion based on gradient elution from reversed-phase chromatography and peptide sequence data. Furthermore, using mass spectrometric analysis of the peptide fraction from SEC corresponding to 750 Da, we were unable to detect amino acid sequences corresponding to plasma proteins.
How can these results and those originating from the Comper group be reconciled? We suggest that 2 pieces of this group's experimental data might need to be reinterpreted: first, the specificity of the Biuret reaction under the conditions used in urine studies by Greive et al15, and second, urine gel filtration results following the intravenous administration of tritiated serum albumin to human subjects23.
The specificity of the Biuret reaction has long been known to be only relative25,26; both keto-acids and free amino acids are known to react with this reagent to give the Biuret chromogen25,26. This lack of specificity is not relevant when applied to plasma, which has a protein content some 3 orders of magnitude greater than that of normal urine. To achieve measurable absorbances with the Biuret method applied to normal urine, the Comper group increased the proportion of analyte volume from less than 1.5% to 20% of the volume of the Biuret reagent. This will also amplify any nonspecific reaction. Indeed, using a mixture of physiologic amounts of nonpeptide urinary constituents, we have generated significant apparent peptide levels. To avoid this interference when the Biuret reagent is used with urine, clinical chemists have acid-precipitated urinary proteins before analysis to remove nonpeptide chromogens27.
Size-exclusion chromatography media permit analyte resolution over a defined mass range. Indeed, even with the use of an agarose-dextran gel having an exclusion limit of approximately 10 kD, we could not always reliably separate free amino acids from small peptides (<5-mer). Osicka et al have presented data based on gel chromatography in which a medium having an exclusion limit of 150 kD was used, and interpreted elution profiles in terms of peptide fragments23,28. Complete calibration data for such chromatographic separations have not been presented. It is, as we have found here, difficult to be sure that free amino acids and small peptides do not cochromatograph on SEC, or even behave anomalously with peptide elution occurring after free amino acids. This suggests a different interpretation of certain results reported by Comper et al: it seems possible that a significant proportion of the radioactivity interpreted as being due to protein fragments may be due to free amino acid.
If there is a pathway for tubular retrieval, partial degradation, and re-excretion (presumably by secretion) for plasma proteins readily filtered at the glomerulus, it seems curious that it is unaffected in the Fanconi syndrome. Our results, particularly in respect of peptide excretion, may reflect the activity of brush-border peptidases on the glomerular filtrate, as well as uptake from the lumen. We have found patterns of protein and peptide excretion in Fanconi syndromes that are fully consistent with the selective filtration and efficient tubular reuptake model. It is, however, of interest that peptides in the range 2 to 5 kD are excreted in 12.9 3.9-fold excess in Fanconi compared with normal urine. Although a variety of peptide hormones, including insulin, has been previously shown to be excreted in large excess, there probably remain many other peptides to be identified to account for the substantial mass of peptides between 2 to 5 kD amounting to 55.5 16.4 mg/24hr excreted by Fanconi patients. Identification of such peptides might suggest the smallest peptide structure in vivo that is processed by the proximal tubular uptake pathway and, therefore, by implication, megalin/cubilin-mediated endocytosis. In contrast to low-molecular-weight peptides, free amino acid excretion in Fanconi was elevated only 2.14 0.73-fold.
CONCLUSION
Our results are most consistent with a model in which any polypeptides >750 Da derived from plasma proteins are excreted in normal human urine at a level less than that of intact proteins. The amino acid analyses and mass spectrometric data reported here appear to be incompatible with previous interpretations of human data from use of the Biuret reaction, and of the size distribution of radioactivity in urine following administration of radiolabeled albumin. We suggest that this difference may be due to lack of specificity of the Biuret reaction and overlapping elution of small peptides and free amino acids in size-exclusion chromatography. In animal models, direct experimental data based on micropuncture or tracking of proteins labeled with fluorophores should be helpful in further unraveling the physiologic pathway(s) of protein handling by the renal tubule.
References
| 1. | D'AMICO G & BAZZI C. Pathophysiology of proteinuria. Kidney Int 2003; 63: 809–825. | PubMed | |
| 2. | KASISKE BL & KEANE WF. Laboratory assessment of renal disease: Clearance, urinalysis and renal biopsy,. inBrenner and Rector's The Kidney 2000; vol 1 6th ed, edited by Brenner BM Philadelphia, W.B. Saunders. |
| 3. | PAVENSTADT H, KRIZ W & KRETZLER M. Cell biology of the glomerular podocyte. Physiol Rev 2003; 83: 253–307. | PubMed | ChemPort | |
| 4. | TOJO A & ENDOU H. Intrarenal handling of proteins in rats using fractional micropuncture technique. Am J Physiol 1992; 263: F601–606. | PubMed | ISI | ChemPort | |
| 5. | MAACK T. Renal handling of proteins and polypeptides,. inHandbook of Physiology, Section 8, Renal Physiology 1992; vol II edited by Windhager E New York, Oxford University Press pp: 2039–2082. |
| 6. | NORDEN AG, LAPSLEY M & LEE PJ et al. Glomerular protein sieving and implications for renal failure in Fanconi syndrome. Kidney Int 2001; 60: 1885–1892. | Article | PubMed | |
| 7. | CHRISTENSEN EI, DEVUYST O & DOM G et al. Loss of chloride channel ClC-5 impairs endocytosis by defective trafficking of megalin and cubilin in kidney proximal tubules. Proc Natl Acad Sci U S A 2003; 100: 8472–8477. | PubMed | |
| 8. | MOESTRUP SK & VERROUST PJ. Megalin- and cubilin-mediated endocytosis of protein-bound vitamins, lipids, and hormones in polarized epithelia. Annu Rev Nutr 2001; 21: 407–428. | Article | PubMed | ISI | ChemPort | |
| 9. | VERROUST PJ, BIRN H & NIELSEN R et al. The tandem endocytic receptors megalin and cubilin are important proteins in renal pathology. Kidney Int 2002; 62: 745–756. | Article | PubMed | |
| 10. | BERGERON M, GOUGOUX A & NO&EUML;L J et al. The renal Fanconi syndrome,. inThe Metabolic and Molecular Basis of Inherited Disease 2001; vol III 8th ed, edited by Scriver CR, Beaudet AL, Sly WS, Valle D New York, McGraw-Hill pp: 5023–5038. |
| 11. | NORDEN AG, SCHEINMAN SJ & DESCHODT-LANCKMAN MM et al. Tubular proteinuria defined by a study of Dent's (CLCN5 mutation) and other tubular diseases. Kidney Int 2000; 57: 240–249. | PubMed | |
| 12. | WRONG OM, NORDEN AG & FEEST TG. Dent's disease: A familial proximal renal tubular syndrome with low-molecular-weight proteinuria, hypercalciuria, nephrocalcinosis, metabolic bone disease, progressive renal failure and a marked male predominance. Q J Med 1994; 87: 473–493. | ISI | ChemPort | |
| 13. | LOWE C, TERREY M & MACLACHLAN E. Organic aciduria, decreased renal ammonia production, hydropthalmos and mental retardation: A clinical entity. Am J Dis Child 1952; 83: 164–184. | ISI | ChemPort | |
| 14. | RUSSO LM, BAKRIS GL & COMPER WD. Renal handling of albumin: A critical review of basic concepts and perspective. Am J Kidney Dis 2002; 39: 899–919. | PubMed | |
| 15. | GREIVE KA, BALAZS ND & COMPER WD. Protein fragments in urine have been considerably underestimated by various protein assays. Clin Chem 2001; 47: 1717–1719. | PubMed | ISI | ChemPort | |
| 16. | CHRISTENSEN EI & BIRN H. Renal handling of albumin in normal rat. Kidney Int 2000; 57: 1207–1209. | PubMed | |
| 17. | GREIVE KA, OSICKA TM & RUSSO LM et al. Fragmentation of filtered proteins and implications for glomerular protein sieving in Fanconi syndrome. Kidney Int 2002; 61: 1549–1550. | PubMed | |
| 18. | CUTILLAS PR, NORDEN AG & CRAMER R et al. Detection and analysis of urinary peptides by on-line liquid chromatography and mass spectrometry: Application to patients with renal Fanconi syndrome. Clin Sci (Lond) 2003; 104: 483–490. | PubMed | |
| 19. | LLOYD SE, PEARCE SH & FISHER SE et al. A common molecular basis for three inherited kidney stone diseases. Nature 1996; 379: 445–449. | Article | PubMed | ISI | ChemPort | |
| 20. | COCKROFT D & GAULT M. Prediction of creatinine clearance from serum creatinine. Nephron 1976; 16: 31–41. | PubMed | ISI | ChemPort | |
| 21. | NORDEN AG, LAPSLEY M & IGARASHI T et al. Urinary megalin deficiency implicates abnormal tubular endocytic function in Fanconi syndrome. J Am Soc Nephrol 2002; 13: 125–133. | PubMed | ISI | ChemPort | |
| 22. | Urine,. inGeigy Scientific Tables 1981; edited by Lentner C Basle, Ciba-Geigy Ltd. pp 62–93. |
| 23. | OSICKA TM, HOULIHAN CA & CHAN JG et al. Albuminuria in patients with type 1 diabetes is directly linked to changes in the lysosome-mediated degradation of albumin during renal passage. Diabetes 2000; 49: 1579–1584. | PubMed | ISI | ChemPort | |
| 24. | SCHOEL B & PFLEIDERER GU. The amount of Tamm-Horsfall-protein in the human kidney, related to its daily excretion. J Clin Chem Clin Biochem 1987; 10: 681–682. |
| 25. | DOETSCH K & GADSDEN RH. Determination of total urinary protein, combining Lowry sensitivity and Biuret specificity. Clin Chem 1973; 19: 1170–1178. | PubMed | |
| 26. | UPDIKE SS & DOUNCE AL. Reaction of histidine with the Biuret reagent in the presence of oxygen. Anal Biochem 1963; 6: 1–12. | PubMed | |
| 27. | HENRY RJ, CANNON DC & WINKLEMAN JW. Clinical Chemistry: Principles and Technics 1974; 2nd ed Hagerstown, Harper & Row. |
| 28. | OSICKA TM, PANAGIOTOPOULOS S & JERUMS G et al. Fractional clearance of albumin is influenced by its degradation during renal passage. Clin Sci (Lond) 1997; 93: 557–564. | PubMed | |
Acknowledgments
This work was funded by The Sir Jules Thorn Charitable Trust. P.R.C. acknowledges support from the St. Peters Trust for Bladder and Kidney Research, and from the Ludwig Institute for Cancer Research. We thank the Biochemical Genetics Unit at Peterborough District Hospital for the measurements of 'physiologic' amino acids. An abstract of this work was presented at the Annual Meeting of the American Society for Nephrology, San Diego, CA, November, 2003.
MORE ARTICLES LIKE THIS
These links to content published by NPG are automatically generated
RESEARCH
Crystal structure of the PI 3-kinase p85 amino-terminal SH2 domain and its phosphopeptide complexes
Nature Structural Biology Article (01 Apr 1996)
Oncogene Original Article
Kidney International Original Article
Protein sequencing by mass analysis of polypeptide ladders after controlled protein hydrolysis
Nature Biotechnology Research (01 Oct 2004)
Kidney International Original Article
