| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
INSERM U402,
2
GRENEMH,
3
Laboratoire de Biochimie A, CHU Saint-Antoine, 75571 Paris Cedex 12, France.
a Address correspondence to this author at: INSERM U402, CHU Saint-Antoine 27, rue de Chaligny, 75571 Paris Cedex 12, France. Fax +33 1.44.75.99.80.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
-aminobutyric acid (7.6%), and for the intermediate concentration
of Asp (8.7%). Between-run precision (CV) was <3.1% for 17 AAs and
<8.0% for 24 of 25 AAs (CV Asp = 12.0%). For retention times,
within-run precision was <0.4% and between-run precision <1.8%.
Excellent relations were found between the results from the Hitachi
L-8500A and the widely used Beckman 6300 analyzer (0.929
r 0.999). The detection was still linear at 5 µmol/L
except for Pro and hydroxyproline (20 µmol/L). The upper limit was at
least 2500 µmol/L for 13 AAs and at least 1000 µmol/L for 27 of 29
AAs (anserine = 500, Val = 600 µmol/L). Values from 100
human plasma samples agreed with previously published data. We conclude
that the results obtained with the Hitachi L-8500A are satisfactory
when compared with those of other AA analyzers utilizing the same
method. Furthermore, the Hitachi L-8500A displays several advantages
including programming flexibility, microsample capacity, low noise
plotting, ammonia filtering, and manual repacking of the analytical
column.
Key Words: indexing terms: ninhydrin reaction • cation-exchange chromatography • analytical performance evaluation
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
The classic ion-exchange separation followed by postcolumn derivatization with ninhydrin has been considerably improved since its initial inception (13), particularly with availability of modern dedicated AA analyzers (14)(15). However, one remaining problem is the relative instability of the ninhydrin reagent, limiting the use of the ninhydrin/acetate buffer mixture to ~2 weeks. This also probably explains why within-run and between-run precisions are so poor for an automated technique (5–10% or more). The design of the new Hitachi L-8500A is therefore of interest. Although the analytical approach is very conventional, AAs being separated with a cation-exchange resin and their concentration determined with the ninhydrin reaction, the instrument provides for extemporaneous automatic mixing of the detection reagent. The Hitachi L-8500A thus allows very accurate detection and full use of the reagent, regardless of how often the apparatus is used.
The aim of this work was to evaluate the analytical performance of this newly available apparatus.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
). The elution program used (Table 2
) was slightly modified from that set by Hitachi to improve the
separation of specific AAs. The separation of Asn, Glu, and Gln was
improved by decreasing the column temperature by 5 °C at 9 min and
by 4 °C at 20 min, that of Gly, Ala, citrulline (Cit), and
-aminobutyric acid (-ABA) by starting the increased flow of
buffer 2 earlier (on at 31.5 min instead of 33.5 min), and that of Val
by starting the 100% buffer 2 step later (on at 44.1 min instead of
43.0 min). An improved separation of Trp and ethanolamine was obtained
by increasing the concentration of benzyl alcohol in buffer 3. Correct
separation of basic AAs was obtained by starting the flow of buffer 4
earlier (on at 74.1 min instead of 77.0 min), adding 10% buffer 2 at
79 min, and decreasing the column temperature by 10 °C at 86 min.
|
|
Detection was by spectrophotometry at 570 and 440 nm with the ninhydrin reaction.
Before the next sample injection, the column was equilibrated with 100% buffer 1 for 19 min (total run time 149 min).
buffers
All buffers can be either purchased (Mitsubishi, Japan) as a whole
package or prepared. We prepared the buffers with HPLC-grade water
generated with a Milli-Q water purification system (Millipore,
Molsheim, France). The chemicals and solvents were of analytical grade.
Trilithium citrate tetrahydrate, lithium chloride, citric acid
monohydrate, lithium hydroxide monohydrate, and benzyl alcohol were
purchased from Merck (Darmstadt, Germany). Pure ethanol was supplied by
Farmitalia Carlo Erba (Milano, Italy), thiodiglycol was purchased from
Prolabo (Paris, France), and caprylic acid from Sigma-Aldrich
(Saint-Quentin Fallavier, France). Chemicals were accurately weighed so
that pH measurement was not required. Each buffer was filtered through
a 0.22-µm filter (Millipore) before use, and kept under nitrogen in
the apparatus. Throughout the elution program the flow rate for buffer
solutions was 0.35 mL/min.
ninhydrin reagent
The ninhydrin reagent, supplied by Wako Pure Chemical Industries
(Osaka, Japan), consisted of 1 L of ninhydrin solution (containing
ninhydrin, sodium borohydrate, and propylene glycol monomethyl ether)
and 1 L of buffer solution (containing dihydrate lithium acetate,
glacial acetic acid, and propylene glycol monomethyl ether). During
sample running, the delivery pump for the ninhydrin reagent
automatically mixes the two solutions kept under nitrogen in the
apparatus. The flow rate for ninhydrin solution was 0.30 mL/min.
calibrator and sample preparation
A 100 µmol/L calibrator stock solution containing 37
physiological AAs was prepared with commercial solutions
(Sigma-Aldrich). Gln (Sigma-Aldrich) was added at the same
concentration and acetyllysine (Aclys; Sigma-Aldrich) was added at 1000
µmol/L.
We evaluated pooled human plasma from hospitalized patients (within-run and between-run precision assays), physiological fluids of various origins such as human or rat plasma, supernatant of homogenate from rat tissue (regression comparison), commercial calibrator solutions (range of linearity), and human plasma from healthy volunteers (determination of physiological range in plasma). In addition, the Bio-Rad urine control solution (Ivry-sur-Seine, France) was used to measure AA concentrations in urine (within-run and between-run precision assays), especially that of 3-methylhistidine (3MH), an interesting AA to study in both malnutrition and hypercatabolic states, given that its urinary excretion reflects the myofibrillar catabolism in muscle(16).
Our procedures complied with the Helsinki Declaration of 1975 for human subjects, as revised in 1983. Animal care complied with the guidelines of our institution, two of us (C.C.-L. and L.C.) being officially authorized (no. 004963 and no. 005226) by the French Ministry of Agriculture and Forestry for animal experimentation.
Blood was drawn into heparinized tubes after a 5-h fast for rats, and either in the postabsorptive state for hospitalized patients or after an overnight fast for 100 healthy volunteers (36 ± 10 years; 41 men, 59 women). Plasma was immediately separated by cold centrifugation (4 °C, 3500g), deproteinized without delay with sulfosalicylic acid (40 g/L), and analyzed or stored at -80 °C until analyzed.
Rat tissues (liver, skeletal muscles) were quickly removed at sacrifice, wiped, weighed, and frozen in liquid nitrogen. Tissue samples were homogenized (4 °C) in 10% trichloroacetic acid (10 mL/g of tissue), containing 0.5 mmol/L EDTA, with an Ultra-Turrax T25 tissue disrupter (Médi Sciences, Saint-Maur-des-Fossés, France). The acid-soluble fraction was separated by cold centrifugation (4 °C, 3500g). Free AA concentrations were measured in the supernatant.
Before analysis, the control and samples were diluted (1:1 by vol) in buffer 1 containing 1000 µmol/L Aclys as internal calibrator.
The within-run precision was studied by performing 10 consecutive runs of the plasma pool from hospitalized patients and of the Bio-Rad urine control solution. To have a wide range of values for plasma, three concentrations were studied: low (L, plasma pool diluted with buffer 1; 1:2 by vol), medium (M, plasma pool), and high (H, plasma pool supplemented with the AA calibrator solution supplied by Sigma; 1:1 by vol). The within-run precision of retention times was studied by 10 consecutive injections of the M concentration plasma pool and that of the urine control solution.
Between-run precision assays were carried out with the M concentration of the plasma pool and the Bio-Rad urine control solution measured in 10 different series. Series were defined by new calibration procedures when the ninhydrin reagent and (or) one or more buffers had to be changed.
A regression comparison was performed by analyzing 77 different physiological fluids [human or rat plasma, supernatant of homogenate from rat tissue (liver and skeletal muscles)] with the Hitachi L-8500A and the Beckman 6300 (Palo Alto, CA), a widely used system for AA analysis.
Linearity and detection limit assays were performed by serial dilutions of the calibration solutions in buffer 1. A Gln solution was diluted in the same manner. The concentrations studied ranged from 5 to 2500 µmol/L.
Sample carryover was analyzed as recommended by the French Society for Clinical Biology (17). Two plasma pools were used: AA at L and H concentrations. We measured the initial concentration for each AA in each specimen (L1 and H1), then the H concentration twice (H2, H3), and then L concentration twice (L2, L3). This sequence (H2, H3, L2, L3) was assayed six times. For each AA we used the Student's paired t-test to compare the mean value obtained for L2 with that obtained for L1, and the mean value of L3 with that of L2.
statistical analysis
Results were expressed as mean ± SD.
Student's paired t-test and regression were applied with the PCSM (Programme Conversationnel de Statistiques pour les Sciences et le Marketing) Deltasoft program (Grenoble-Meylan, France). Differences were considered significant when P <0.05.
The Kolmogorov–Smirnov test showed that human plasma AA concentrations were normally distributed, in agreement with the literature (18).
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
. Under these
conditions, the Hitachi L-8500A apparatus separated 38 AAs (including
Aclys as internal calibrator) during the elution program.
Typical chromatograms for the calibration solution and for human plasma
are shown in Fig. 1
. AA separation is satisfactory except for that of Asn and Glu,
a major concern of AA separation by ion-exchange chromatography.
|
performances
The results of the within-run precision assays for concentrations
and retention times performed on human plasma pools are presented in
Table 3
. For concentrations, the CVs were <4.5% for L values of 24 of
25 AAs (CV -ABA = 7.6%). CVs were <3.8% for M values, except for
Asp (8.7%), and <1.9% for H values for all the AAs studied.
Retention times were very stable, with CVs from 0.0% to 0.3%. These
CVs appeared better than those shown by other HPLC systems with the
same (14) or other
(8)(9)(12) methods. The CVs of the
concentrations in the urine samples (data not shown) were about twice
those observed with serum, <4.0% for most of the AAs (20 of 23).
Three AAs had high CVs: -ABA (6.9%), Pro (14.4%), and Cit
(19.8%). This may be explained by the difficulty in measuring AAs at
low concentrations in urine samples (23, 25, and 12 µmol/L,
respectively). Values for 3MH were 108 ± 3 µmol/L (CV =
2.7%). The results of the within-run precision assay for the retention
time of AAs from the urine sample were similar to those obtained with
serum.
|
The results of between-run precision assays for concentrations and
retention times performed on a human plasma pool are presented in Table 4
. Concerning between-run precision of the peak area, most of the
AAs studied (19 of 25) had CVs <5.2%. The CVs of 5 AAs (Glu, Cit,
-ABA, Trp, and Arg) were between 5.6% and 7.9%. The comparatively
poor result for Asp (CV = 12.0%) may have been due to its low
concentration (7 µmol/L) in the sample used.
|
Concerning the between-run precision of peak position, CVs of 21 AAs
were <1.0%. The CVs of four AAs (Ser, Asn, Glu, and Gln), with
retention times between 17.62 and 23.36 min, were between 1.1% and
1.8%. These poorer results could have occurred because two buffers
were programmed simultaneously in different proportions. Hence small
changes in composition occurring during preparation could have caused
additional variations on the chromatographic profile observed. CVs of
the concentrations of the urine samples (data not shown) were higher
than those observed with serum, <10% for 17 of 24 AAs. The
particularly poor results for -ABA (15.3%), Asp (19.7%), and Glu
(23.1%) were probably due to their low concentration—20, 12, and 26
µmol/L, respectively. Values for 3MH were 108 ± 6 µmol/L
(CV = 5.6%). The results of the between-run precision assay for
the retention time of AAs from the urine sample were similar to those
obtained with serum (CV <1.6%).
Collectively, the results of the between-run precision assays were satisfactory and close to or even better than those obtained by classic ion-exchange chromatography (14)(19), other HPLC techniques (8)(9)(10)(12), or gas chromatography (20).
Data for the relation between the results of the two systems used
(Hitachi L-8500A, Beckman 6300) are presented in Table 5
. The n values were different for each AA because when they were
analyzed with the Beckman 6300, the samples were assayed under routine
conditions for the apparatus (diluted 1:6 by vol vs 1:1 by vol with the
Hitachi L-8500A); thus concentrations of some AAs were less than the
lower detection limit of the Beckman 6300. Also, for some samples, some
AAs were not measurable with either apparatus. A fair relation was
found for all the AAs analyzed (21), r values being >0.929.
Furthermore, for a large majority of the AAs studied (15 of 21),
r was >0.980. Asp was not included in this analysis; its
measurement on the Beckman 6300 was not accurate because of its
coelution with reduced glutathione when analyzing rat tissues
(Coudray-Lucas et al., unpublished data).
|
The results of the linearity limit assays showed that the reaction was
linear up to a concentration of at least 2500 µmol/L for 13 AAs: Asp,
Thr, Ser, Glu, Gln, Pro, Gly, Ala, -ABA, Cys, Ile, Leu, and Tyr. The
reaction was linear up to 2000 µmol/L for Arg; 1500 µmol/L for
phosphoserine, taurine, cystathionine, and Phe; 1250 µmol/L for Met
and 1-methylhistidine; and 1000 µmol/L for Asn, hydroxyproline
(HyPro), ornithine, Lys, His, 3MH, and carnosine. The upper limits of
linearity for Val and anserine were 600 and 500 µmol/L, respectively.
However, limit values were always higher than usual plasma
concentrations for all AAs. The lower limit of detection was <5
µmol/L for all the AAs except for Pro and HyPro (20 µmol/L).
However, we did not measure concentrations between 5 and 20 µmol/L.
Further studies are therefore required for more accurate determination
of the lower detection limit for these last two AAs.
Concerning sample carryover analysis, the comparison for each AA between its mean concentration in the plasma pool of low concentration, assayed after that of high concentration (L2) and (a) its initial concentration (L1) and (b) its mean concentration determined after analyzing L2 (i.e., L3), showed no statistical differences (data not shown), indicating no contamination between samples.
physiological range for aa concentrations in human plasma
The results of the analysis of plasma samples from 100 healthy
volunteers are presented in Table 6
. These results were very close to those obtained by other HPLC
systems with the same (1)(2)(21)
or other (4)(5)(9) methods.
However, the results for Asn and Glu concentrations obtained in the
present study were moderately lower than those in the literature. This
may partly be explained by the difficulty in obtaining a correct
separation of these two AAs with the ion-exchange chromatography
system. Furthermore, in our study, the mean concentration for Glu was
specifically low, but the total of Glu plus Gln was in agreement with
other studies (4)(5)(21). High Glu
concentrations, as reported in other publications, may be due to the
partial conversion of Gln, which is unstable in solution, to Glu during
storage of the samples. In our study, we analyzed the samples very
promptly after blood collection, so that, as expected, the
concentrations of Gln we measured were high and those of Glu low.
|
In conclusion, the Hitachi L-8500A affords accurate results for physiological AA analysis.
This apparatus displays some further advantages:
• During the analysis program, buffer solution switching times and proportions, together with column temperature (from 30 to 70 °C in 1 °C steps), can be modified as required. Similarly, the setting up of new programs is very simple.
• The possibility of analyzing small quantities of samples with microvials. In this case, only 32 µL are necessary when using the recommended 20-µL injection option (it is possible to inject 10 µL, reducing the volume required to 22 µL).
• A refrigerated (4 ± 2 °C) sample rack (60 samples), allowing the conservation of samples in appropriate conditions.
• Two plungers for each pump (two pump heads), reducing noise and giving a very linear stable baseline, allowing an average limit of sensitivity of 10 pmol (for a signal-to-noise ratio of 3).
• A loop that leads the selected buffer to the required temperature before its run onto the analytical column, also reducing noise.
• An ammonia filter column.
• The possibility to manually treat and pack the analytical resin. After resin treatment (acetone, ethanol, 2 mol/L hydrochloric acid, and 2 mol/L lithium hydroxide), the column is automatically filled (1 h, 2 x 107 Pa). This procedure is required about every 800 runs. This prevents the exchange of the column, allowing the same accurate determination with time.
• Failures are infrequent, provided the apparatus is regularly maintained. Special attention should be paid to cleaning the column filters and changing pump seals on a monthly basis. During our 28-month experimental period, the apparatus was idle for ~42 days because of equipment failures or having to wait for materials, and 25 days were spent on maintenance. Allowing for this, ~2000 samples were analyzed per year at a cost of about $15 (US) per analysis (excluding labor costs), similar to the cost calculated for the Beckman 6300.
|
Footnotes |
|---|
-ABA, -aminobutyric acid; Cit, citrulline; Aclys, acetyllysine; 3MH, 3-methylhistidine; and HyPro, hydroxyproline. 
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. Raynaud-Simon, M. Kuhn, J. Moulis, J. Marc, L. Cynober, and C. Loi Tolerance and Efficacy of a New Enteral Formula Specifically Designed for Elderly Persons: An Experimental Study in the Aged Rat J. Gerontol. A Biol. Sci. Med. Sci., July 1, 2008; 63(7): 669 - 677. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Jourdan, L. Cynober, C. Moinard, M. C. Blanc, N. Neveux, J. P. De Bandt, and C. Aussel Splanchnic sequestration of amino acids in aged rats: in vivo and ex vivo experiments using a model of isolated perfused liver Am J Physiol Regulatory Integrative Comp Physiol, March 1, 2008; 294(3): R748 - R755. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Rouge, C. Des Robert, A. Robins, O. Le Bacquer, C. Volteau, M.-F. De La Cochetiere, and D. Darmaun Manipulation of citrulline availability in humans Am J Physiol Gastrointest Liver Physiol, November 1, 2007; 293(5): G1061 - G1067. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Parlesak, I. Negrier, N. Neveux, C. Bode, and L. Cynober Arginine Does Not Exacerbate Markers of Inflammation in Cocultures of Human Enterocytes and Leukocytes J. Nutr., January 1, 2007; 137(1): 106 - 111. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Osowska, T. Duchemann, S. Walrand, A. Paillard, Y. Boirie, L. Cynober, and C. Moinard Citrulline modulates muscle protein metabolism in old malnourished rats Am J Physiol Endocrinol Metab, September 1, 2006; 291(3): E582 - E586. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Phanvijhitsiri, M. W. Musch, M. J. Ropeleski, and E. B. Chang Heat induction of heat shock protein 25 requires cellular glutamine in intestinal epithelial cells Am J Physiol Cell Physiol, August 1, 2006; 291(2): C290 - C299. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B.-M. Iresjo, U. Korner, B. Larsson, B.-A. Henriksson, and K. Lundholm Appearance of Individual Amino Acid Concentrations in Arterial Blood During Steady-State Infusions of Different Amino Acid Formulations to ICU Patients in Support of Whole-Body Protein Metabolism JPEN J Parenter Enteral Nutr, July 1, 2006; 30(4): 277 - 285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Osowska, C Moinard, C Loi, N Neveux, and L Cynober Citrulline increases arginine pools and restores nitrogen balance after massive intestinal resection Gut, December 1, 2004; 53(12): 1781 - 1786. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-C. Preiser, J. Berre, A. Van Gossum, L. Cynober, B. Vray, Y. Carpentier, and J.-L. Vincent Metabolic Effects of Arginine Addition to the Enteral Feeding of Critically III Patients JPEN J Parenter Enteral Nutr, July 1, 2001; 25(4): 182 - 187. [Abstract] [PDF]
|
|
|
|
 |
|
 R. Venta Year-Long Validation Study and Reference Values for Urinary Amino Acids Using a Reversed-Phase HPLC Method Clin. Chem., March 1, 2001; 47(3): 575 - 583. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Schlegel, C. Coudray-Lucas, F. Barbut, J. Le Boucher, A. Jardel, S. Zarrabian, and L. Cynober Bacterial Dissemination and Metabolic Changes in Rats Induced by Endotoxemia following Intestinal E. coli Overgrowth Are Reduced by Ornithine {{alpha}}-Ketoglutarate Administration J. Nutr., December 1, 2000; 130(12): 2897 - 2902. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
