A Compound from Uremic Plasma and from ... - Clinical Chemistry

2 downloads 0 Views 765KB Size Report
A Compound from Uremic Plasma and from Normal Urine Isolated by Liquid. Chromatography and Identified by Nuclear Magnetic Resonance. P. GallIce,' J. P. ...
CLIN. CHEM. 31/11, 30-34 (1985)

A Compound from Uremic Plasma and from Normal Urine Isolated by Liquid Chromatography and Identified by Nuclear Magnetic Resonance P. GallIce,’ J. P. Monti,1 A. Crevat,”3 C. Durand, and A. Murlsasco2 present in normal urine and in ultrafiltrates of uremic plasma in the fraction of so-called “uremic middle molecules” was isolated by liquid chromatography. Preliminary studies, including amino acid analysis, characterization of uronic acids, and ultraviolet spectroscopy, show that the molecule contains glycine, a uronic acid, and an aromatic ring. Characterization by 1H and ‘3C nuclear magnetic resonance spectrometry shows conclusively that this compound

Then, to eliminate the TBAP, we passed the solution through a 1 x 30 cm glass column packed with Sephadex DEAE A-25 (Pharmacia, Uppsala, Sweden) equilibrated with ‘Fris HC1 buffer (10 mmolJL, pH 8.6).The column is washed for 3 h at a flow rate of 20 mL/h with the same buffer,to remove the unretained TBAP. The 2-5-10 peak is eluted with a gradient of NaC1 (0 to 0.15 mol/L) in the Tris HC1 buffer (3). The eluate ismonitored by absorbance

acid, which has been previously described by Zimmerman et al., using quite different techniques of isolation and identification (C/in Nephro/14: 107, 1980; FEBSLett129: 237, 1981).

TBAP

A compound

is a double conjugate of glucuronid#{225}te-o-hydroxyhippuric at 254 nm and collectedin a fractioncollector.We verified

AdditIonal Keyphrases: organic acids

glucuronidafe-o-hydroxyhippuric

acid

Uremic middle molecules (UMM)4 are Mr 300-2000 compounds that accumulate in uremic plasma (1,2). Because of our previous finding that these substances were also present in the urine of normal subjects, we concluded that these UMM had a possible physiological significance (3). So far, various mixtures of these compounds have shown effects on different biological systems (4-7). A better knowledge of these molecules should therefore be of great interest. Here we describe the isolation of one of these molecules and its identification in particular,

by means of several physicochemical methods; we can now specifr its spatial structure.

Materials and Methods Isolation of the compound. We separated the UMM subfractions as described elsewhere (3): Gel permeation of uremic plasma ultrafiltrate or normal urines on Sephadex G 15 yields crude UMM in fraction2. Next, chromatography by anion exchange on Sephadex DEAE A-25 separates fraction 2 into seven components (2-1 through 2-7). Finally, fraction 2-5 is divided into 10 subfractions (2-5-1 through 25-10) by means “high-performance” liquidchromatography (HPLC), by a reversed-phase ion-pairingprocedure involving tetrabutylammonium phosphate (TBAP) as a counterion (3).

We slightly modified the HPLC technique to a semipreparative method: we inject 200-iL volumes of a 4 g/L solution of subfraction 2-5 and elute, with a concentration of methanol in the mobile phase of 200 mL/L. Purification of the 2-5-10 subfraction. After evaporating the sample under reduced pressure at 37 #{176}C, with a RotaVapor, we dissolved the residue in doubly distilled water. ‘Laboratoire de Biophysique, Facult#{233} de Pharmacie, 27 Ed. J. Moulin 13385, Marseille cedex 5, France. tCentre d’H#{233}modialyse, Hopital Sainte Marguerite, Marseille, France. 3Address correspondence to this author. Nonstandard abbreviations: UMM, uremic middle molecules; HPLC, “high-performance” liquid chromatography; TBAP, tetrabutylammonium phosphate, NMR, nuclear magnetic resonance. Received March 20, 1984; accepted September 14, 1984. 30 CLINICAL CHEMISTRY, Vol.31, No. 1,1985

removal by the method of Slonecker et al. (8). We desalted the 2-5-10 fraction by gel permeation chromatography in a 2.5 x 100 cm glass column packed with Sephadex G 15, eluting with doubly distilled water. The subfraction thus isolated (as its Na salt) is stored lyophilized. Identification of the 2-5-10 subfraction. We characterized uronic acids by Dische’s technique (9). Ultraviolet spectra were recorded with a Uvikon 820 spectrophotometer (Roche Bioelectronic Kontron). We determined the amino acid composition of the sample with a continuous-flow CarloErba type 3A-27 analyzer, after the sample was hydrolyzed for18 h in a sealed ampoule at 110 #{176}C with 6 molJL HC1 (10). We hydrolyzed the 2-5-10 compound enzymically with f3glucuromdase, according to Greenblatt et al.(11). We recorded the ‘H and 13C NMR spectra at 200.13 and 50.32 MHz, respectively,with a Bruker AM 200 spectrometer (Centre Universitaire de RMN, Faculte de Pharmacie, Marseille) in the pulsed Fourier transform mode. Typical experimental conditions for ‘H NMR were: sweep width 2400 Hz, pulse width 4 zs, acquisitiontime 3.4 s,number of scans 20-100. The 2-5-10 subfraction was placed in a 5-mm (o.d.) tube and studied at 20.0 (±0.5) #{176}C in 2H20 solution (--2-3 mg/0.5 mL, pH 7.1 ± 0.1),with the solvent forming the internal lock. ‘H chemical shiftsare measured with

reference to the resonance of 2,2-dimethyl-2-silapentane-5sulfonate [3-(trimethylsilyl)-1-propanesulfonate] in 2H20 packed in a concentric capillary tube. Proton chemical shift assignments are classicallyperformed by spectrum integrations,homonuclear double irradiation,and comparison with unconjugated compounds previously recorded under the same conditions: for comparison we used salicylic acid, glycine, n-glucuronic acid, and o-hydroxyhippuric acid (Sigma Chemical Co.). Typical experimental conditions for ‘3C NMR were: sweep width 11 900 Hz, pulse width 9 s, acquisitiontime 0.69 s, number of scans 20 000-70 000. The 2-5-10 subfraction, placed in a 10 mm (o.d.)tube, was studied at 35.0 (±0.5) #{176}C in 2H20 solution (-2-3 mg/2 mL) with the solvent as the internal lock.‘3C chemical shifts are measured with reference to tetramethylsilane. ‘3C resonances were obtained by

comparison with results for unconjugated compounds previously recorded under the same conditions. ‘3C chemical shift assignments for salicylic acid, D-glucuronic acid, and ohydroxyhippuric acid were determined by comparison with analogous compounds (12, 13) and by considering the addstivityof substituent effects(14). For spectrum simulation, we used a computer program developed for an Apple H computer 64K (M. Sarrazin, unpublished), derived from a

IAOCOON program

(Centre de calcul de St J#{233}r#{244}me, Facult#{233}

des Sciences Marseille).

Results Figure 1 shows the results obtained during the different steps of separation. Ion-exchange chromatography after the HPLC isolation step yields a single symmetrical peak (elution volume = 155 ± 10 mL, mean ± SD),corresponding to the 2-5-10 subfraction with no TBAP present. This subfraction tested positive for uronic acids. The ultraviolet spectrum of this subfraction showed three maxima (Figure 2), suggesting the presence of an aromatic ring (230 and 280 inn) and an amide linkage (202 nm) (15). After acid hydrolysis, we detected only glycine, and no other amino acid. Table 1 lists all the ‘H chemical shifts and the coupling constants obtained for the 2-5-10 compound and the four free comparison model molecules: salicylic acid, glycine, ohydroxyhippuric acid, and f3-n-glucuronic acid. Figure 3 shows the 1H NMR spectrum of the 2-5-10 compound and its theoretical spectrum. Table 2 lists all the 13Cchemical shifts obtained for the 2-

-

200

Fig. 2.

Ultraviolet

o

-j---

-.-------

250

300

350

,Unm)

spectrum of 2-5-10 subtraction compound

tfl,allll,.le

No,m&u,in.

01 u,#{149}mlCPl.smg

/ G,l

pu,meal.o.n

)OEAE

PeHOtmonCe

*25)

Sop0000. 2-5

F,.*,ion Higt

Liquid

ly reported. After hydrolysis with p-glucuronidase, the ‘H spectra of the hydrolysate showed the spectral features of two of the model molecules: o-hydroxyhippuric acid and /3-D-glucuronic acid.

2

F,.c,io., *fl,OnICencflenOe

5-10 compound as well as those for model products previous-

C’S)

(Sepfladeo

CoomaI0gf

SOnY

Discussion 3

0

R.I.ol00

2-5-

SuDlsactiop

Ion

e

limO

(mi,)

,0

.CflenQe’1

Sop000000

251

I

lol\

JI_ 50

00

50

200 SubI,.nlloo

2-5-

Elu,iOe,OkIm.(I”l)

0

p,,m.eflo,

(Sepheden

G15)

In our previous papers (3,16) we showed that crude UMM from uremic plasma and normal urines have the same qualitative composition. The present results confirm this assertion: the spectral characteristics of the 2-5-10 sample are the same regardless of its origin. Thus we can disregard the possibility that an artefact, e.g., from drugs taken by the patients, eluted with UMM during the chromatography (17), and we can assert that the 2-5-10 compound is an endogenous substance that is present in very low concentrations in normal plasma. The first difficulty in this procedure was to eliminate the counterion (TBAP) after isolation by HPLC. The anionexchange and gel-permeation procedures completely removed the TBAP and the other salts used during the isolation steps. Preliminary results with ‘H NMR spectra showed that this molecule is formed from three components: a uronic acid, glycine, and an aromatic compound. A more detailed study of ‘H NMR spectrum showed that the uronic acid characterized by Dische’s method (9) was f3-r-glucuronic acid. Indeed, if ‘H resonances were produced by the same chemical shifts ranges, whatever the uronic acids, the proton coupling constants, 3J, would be different. Only f3-rglucuronic acid presents an equilibrium Jax,,, Jeqeq for all the protons that produce J values in the range 7-10 Hz (Table 1). By contrast, all other /3-uronic acids have at least one equilibrium J J,, with J values in the range 24 Hz (18). Moreover, the high J,.2. value (7.6 Hz) is in good agreement with the /3-anomer form only (J ‘2’ anomer = 3.7 Hz, Table 1). The deshielding observed about the protons (particularly H,’ and H2’) of /3-glucuronic acid in the 2-5-10 sample with regard to the free compound may be due to ring current from a nearby aromatic component. Moreover, the ‘3C chemical shifts of the f3-r-glucuronic acid ring are analogous to those of the free compound except for C,’ (Table 2). The deshielda

00*.,,,.

(mU

Fig. 1. Chromatographic isolation of 2-5-10 subtraction summarized

CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985

31

Table 1. Proton Chemical Shift Assignments (#{244}) and Coupling Constants (J) C0-NH-CH2-COONa

COONa H4

H4

OH

TC)I

H2N-CH2-COOH H3y

H1

H4.k

OH

( :)I

COOH 0

H H3.

H2

H2 SaIlcyclic acid, Na salt Protons 1 2 3 4

5, ppm

6.75 7.31 6.77 7.65

o.Hydroxyhlppuricacid (Na saft)

Glycine

J, Hz J12=8.2 J13=0.7

5,

ppm

5,

ppm 6.77 7.28

J23=8.2

6.79

J24=1.8

7.56

J, Hz J12=8.3 J13=0.7 J23=8.3 J24=1.7

J.,=8.3

5,

2-5-10 compound, Na salt

ppm

5,ppm

J, Hz

H2

Hi

1

J,Hz J12=8.3 J13=0.5

7.05

J238.2 J=7.8

(A) 3.78 (B) 3.82

.

1’ 2’ 3’ 4’ 5’ For model compound o-glucuronic acid, only proton

7.15 7.37 7.59

J=7.8 3.75

3.27

CH2

H4

p-o-Glucuronlcacld

JAO=-17.3

4.46 3.06

J1.2.=7.9

5.00

J1.2.=7.6

J2.3.=9.5

3.57

3.29 3.35 3.77

J3.4.=10.0 J4.5.=9.7

3.44 3.45 3.76

J2.3.=9.5 J3..=9.0

of the a-anomer can be accuratelydistinguished: 3

=

5.03 ppm and J1‘2

J4.5.=9.7 3.7 Hz.

A

H3

GIy (41’

(45’

II

(42’

B

H4’H3’

II

- ,j 7.8

7.6

7.4

7.2

7.0

ppm

50

4.8

I

3.8

3.6

Fig. 3. 1H NMR spectra of the 2-5-10 compound, as measured experimentally and simulated from theory The simulatedspectra(lowerspectrumof eachpair)are preparedfromthe valueslistedin Table 1 anda 1.5-Hzwidthof the peaks at half-height. (verticalscale x 2); B, glycine and -o-glucuronic acid region ing of the latter results from the influence of the substituent (aromatic compound) on the C,’ hydroxyl oxygen, as reported elsewhere (13). Finally, the Na salt form of the 2-5-10 compound displays a classical deshielding for carbonyl carbon (19). From all of this we conclude that the 2-5-10 product is a conjugate of /3-glucuromc acid (Figure 4). Concerning the aromatic and glycine components, the characteristic coupling features of H2 and H4 suggest the presence of salicylic acid (Table 1), a compound that is likely to be conjugated to glycine. Because glycine is the only amino acid detected in the 2-5-10 compound (20), we used ohydroxyhippuric acid as a model. The coupling features and

32 CLINICAL CHEMISTRY, Vol. 31, No. 1, 1985

3.4

A,

pp

Salic1ic acid region

integration of this compound are identical in 2-5-10 middle molecule except for the glycine AB system (see below). The H, and H3 deshielding in the unknown compound would be induced by spatial proximity of /3-n-glucuronic acid, especially the H,, and by the covalent bond of the o-hydroxyhippuric acid hydroxyl group. This hypothesis is perfectly supported by the ‘3C results (Table2).The passage from ionic (o-hydroxyhippuric acid) to covalent binding (2-5-10) modifies the electronic charges on the ring carbons, giving rise to a classic case of ortho-carbon shielding and para-carbon deshielding. Similarly, the carbon linked to the hydroxyl group is shielded (14). The

Table 2. 13Carbon Chemical Shift Assignments (ppm) 7 8 9 CO-NH-CH2-COONa

COONa >OH 61

(

12

Carbon atoms 1(8) 1 2 3 4 5 6 7

ii-o

6