This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Acetonitrile and methanol used were HPLC grade and were purchased from Carlo Erba. MilliQ water was generated by a milliQ water purification system (Millipore, Molsheim, France). Formic acid, solution of ammonium hydroxide and of trichloroacetic acid, Î²-N-methylamino-L-alanine hydrochloride (BMAA) and 2,4-diaminobutyric acid dihydrochloride (DAB) were provided from Sigma Aldrich (Saint-Quentin Fallavier, France). 2,4 diaminobutyric acid-2, 4, 4-d3 dihydrochloride (D3-DAB) was purchased from CDN isotopes (Pointe-Claire, Canada).
Stock solutions (100 µg.ml-1) for the three standards molecules (BMAA, DAB and D3-DAB) were prepared in water and were stored at 4°C until further dilution were made.
2.2.) Sample material
Samples of Seine river water was collected in April and May 2012 on the "Ile aux cygnes" in Paris and was stored at 4°C until analysis were performed. Samples of biofilm were collected in August 2008 Castelbouc (Voir Annick) and were lyophilized and stored at -20°C until used. PCC 6506 was an axenic strain of cyanobacteria and was obtained from the Pasteur Culture Collection of Cyanobacteria (PCC), Paris, France. This cyanobacterial strain was grown as previously described  and was lyophilized and stored at -20°C until used.
For validation procedure, stock standards were diluted in water each week for obtained calibration (1 µg.ml-1) and validation (0.1 µg.ml-1) working solutions containing both BMAA and D3-DAB.
2.3) Samples preparation:
2.3.1) Calibration samples:
At first, calibration samples were prepared in duplicate at 4 levels of concentrations (5, 25, 100 and 200 ng.ml-1) replicated on 5 different days by diluting the calibration working solution in mobile phase (ACN 0.1% formic acid, water 0.1% formic acid, 65/35, v/v). But after optimization of the regression mode in pure water, we prepared in duplicate only 2 concentrations levels (25 and 200 ng.ml-1) replicated in 5 different days.
2.3.2) Validation samples:
25 mg of dried biofilm or 10 mg of dried cyanobacteria were weighted in a 15 ml centrifuge tube. The sample was then spiked with a determined volume of validation working solution and TCA 0.1M was added to have a final volume of 3 ml. The tube was vortex mixed during 1 minute and then was placed on an ultrasonic bath during 10 minutes. The sample was centrifuged at 4000 rpm for 10 min at room temperature and the supernatant was saved.
Pure water and Seine river water were acidified with formic acid for reaching a pH of 3. 3 ml of water was withdrawn and was spiked with a known volume of validation working solution.
Validation samples were prepared in duplicate or triplicate at 4 levels of concentration ranging from 0.5 to 13 ng.ml-1 for pure water and from 1 to 13 ng.ml-1 for Seine river water, biofilm and cyanobacteria. Five series of samples were prepared alternately by 2 different operators for each matrix.
2.3.3) Solid-Phase-Extraction (SPE) process:
Polymeric cation exchange sorbants, Oasis-MCX (3cc, 60mg, Waters, Guyancourt, France) were used to clean the validation samples. The Oasis-MCX cartridge was previously reported as an effective clean-up for cyanobacteria samples before analyzing BMAA by LC/MS-MS  . In this study, we have fully automatized the procedure described previously using the GX271 liquid handler system (Gilson, Villiers le Bel, France). The condition step involved activating the cartridge (3ml) with 2 ml of methanol followed by 1 ml of water at pH = 3. Once the validation sample had passed through the cartridge, it was rinsed with 2 ml of HCl 0.1M and with 1ml of methanol. Elution was performed using 3ml of methanol basified with 5% of NH4OH. Finally eluents passed through this cartridge were dried under nitrogen and then reconstituted in 200 µl of a mixture ACN containing 0.1% formic acid, water containing 0.1% formic acid (65/35, v/v). The reconstitute sample was transferred to an auto-sampler vial and 5 µl were injected into the LC/MS-MS system.
2.4) Sample analysis by LC/MS-MS:
LC/MS-MS was performed using a liquid chromatography (UltiMate 3000®, Thermo scientific, Illkirch, France) coupled to a TSQ Quantum Access MAX Triple Stage Quadrupole Mass Spectrometer (Thermo scientific, Illkirch, France) equipped with a heated electrospray ionization source (HESI). Chromatographic separations were conducted on a Zic-Hilic column, 3.5 Î¼m, 150 x 2.1 mm (Merck Sequant AB, Umeå, Sweden). This type of column was particularly designed for polar compound and was already described for allowing a good retention of underivatized BMAA  . Mobile phases were acetonitrile containing 0.1% (v/v) formic acid (A) and MilliQ water containing 0.1% (v/v) formic acid (B). We used a linear gradient from 35 to 45% of B in 25 minutes, hold for 2 minutes, ramp over 3 minutes to 35% B and hold to 15 minutes to equilibrate the system. The flow rate was set at 0.2ml/min. Each validation or calibration sample was injected twice in LC/MS-MS system.
MS was operated in positive ion mode with MRM detection using an electrospray voltage of 3000 V, a capillary voltage of 35 V, a tube lens offset of 60 V and capillary and vaporizer temperature were set respectively at 270°C and at 50°C. We used nitrogen as desolvatation gas and argon as collision gas at a pressure of 1.5 mTorr. Three transitions including 119>102, 119>101 and 122>104 were monitored at collision energy of 12V for the two first and 10V for the last. These transitions were used to optimize chromatographic conditions and to quantify BMAA, DAB and D3-DAB respectively. The instrument was also operated in full scan (Q1) during optimization and validation studies in order to visualize interferences that co-elute with BMAA and DAB and matrix effects. For validation samples, MS data acquisition started at 10 minutes after sample injection, and the divert valve was set to waste until data acquisition started to avoid polluting the HESI source of mass spectrometer.
3) Results and discussion
3.1) Optimization of the analytical procedure:
3.1.1) LC/MS-MS analysis
For optimization of the MS-MS detection, a mix of BMAA, DAB and D3-DAB standards was directly infused at a concentration of 50 µg.ml-1. Fragmentation spectra for BMAA and its natural isomer the DAB (figure 1) shows that they have very similar pattern of fragmentation.
According to these patterns, quantification of BMAA and D3-DAB was performed in MRM mode using the specifics transitions 119>102 and 122>104 to quantify respectively BMAA and D3-DAB. Transition 119>101 was also monitored in validation samples for avoid false positive response link to possible presence of DAB in natural samples.
A gradient elution was developed on Zic-Hilic column to provide good chromatographic separation of BMAA and DAB. MRM chromatogram was shown on the left column of Figure 2. Calibration curves for BMAA, DAB and D3-DAB over the concentration range of 0.003 to 5 µg.ml-1 show good linearity. Detection limits (S/N = 3) and quantification limits (S/N = 10) are respectively of 1 and 3 ng.ml-1 with an injection volume of 5 µl for the three compounds (respectively 5 and 15 pg injected on the column).
3.1.2) Sample treatment and SPE procedure
Clean up sample prior to LC/MS-MS analysis reduce interferences that cause matrix effects and increase sensibility of detection particularly in natural sample which contain numerous low molecular mass compounds. Our sample clean up procedure is based on those already published by Li et al in 2012 . However we did slight modifications to increase efficiency and/or to facilitate the implementation of the procedure for treatment of numerous real samples. For example, we equilibrate the system with acidified water, which decreased the loss of BMAA during percolation. For cons, the last step of elution with 5% NH4OH in water was replaced by an elution in 5% NH4OH methanol. This step is less efficient and we had to increase the volume of elution from 2 to 3 ml. But it's around 2 times faster to evaporate 3 ml of methanol than 2 ml of water. This evaporation steps after elution on SPE cartridge was particularly crucial. Indeed it allowed us to concentrate and to reconstitute sample in a weaker solvent which is more consistent with Zic-Hilic column. This two factors take together led to limit loading effect and to have a good reproducibility in chromatographic separation.
3.1.3) Determination of recovery rates
Yield of the different steps of our analytical procedure were monitored by comparing signals of samples spiked before and after sample treatment and SPE procedure. Results are summarized in Table 1. Even after clean-up on SPE, environmental samples extract remained complexes and contained many molecules that were co-eluted with BMAA. These numerous interferents molecules decreases the signal in MRM and increases retention time (MRM chromatogram for cyanobacteria was shown as example on right column of Figure 2). During the optimization of the LC/MS-MS analysis, preliminary studies have demonstrated that the matrix effect was similar for BMAA and D3-DAB in our different environmental matrix (data not shown). Matrix effects vary from 30% in Seine river water from XX% in the cyanobacteria (Table 1). This method was selective, indeed even at a concentration of 25 ng.ml-1 in extract of cyanobacteria cleaned up on SPE, specific transitions for BMAA and D3-DAB can be distinguished from background noise and can be used to quantify the both molecules (Figure 2).
3.2) Validation by total error approach:
3.2.1) Method validation
Validation should ensure suitability between analytical method and specifications for which it was developed. We used a statistical approach based on the total error measurements  , this approach takes into account systematic and random errors committed throughout our analytical procedure. Acceptance criteria would ensure that a determined proportion (say Î²) of future measures lie within acceptance limits, with a high degree of confidence. For this study, the acceptance limits was settled at ± 20% and the minimum probability Î² at 90%.
3.2.2) Corrective factor:
As we shown below, during LC separation, we can easily discriminate between DAB, from its natural isomer BMAA and using specific transition during MS-MS analysis allowed us to discriminate between D3-DAB and DAB. Moreover, behavior of D3-DAB was closest to that of BMAA during sample treatment, SPE procedure and LC/MS-MS analysis. In consequence, we used D3-DAB to monitor the global yield of our analytical procedure (included sample treatment with TCA, SPE cleaning up procedure and matrix effect during LC/MS-MS analysis). This global yield was calculated for D3-DAB in each validation samples and for each matrix. Then the corrective factor obtained was used to correct the measured concentration of BMAA in the sample injected in LC/MS-MS and led to the back-calculated concentration in spiked validation sample before any treatment.
3.2.3) Response function:
We tested four different models of regression for calibration curves and monitored their effect on precision of back calculated concentration obtained for validation samples in pure water. In order to monitored precision, two sets of values were calculated: repeatability and time-different intermediate precision. The relative deviation standards (RSD) of these 2 parameters for each regression models tested were shown on Table 2. We have chosen the best regression model in order to minimize deviation standard in regression parameters of curve calibration and to obtain the best accuracies covering the entire dosing range. For the three highest concentration levels, the four models give similar results in term of repeatability and intermediate precision (ranging from 8 to 12% according to the concentration levels). However for the lowest concentration level (0.5 ng.ml-1), the linear regression through zero model with 2 or 4 points of calibration give better repeatability and intermediate precision than the two others. These both models give similar results in term of precision. So we use linear regression through zero with only two concentrations levels for calibration in order to facilitate the implementation of the analytical procedure in environmental matrices. Nevertheless, even with this model, considerable variability was seen at the lowest concentration level in validation samples.
3.2.3) Trueness, precision and accuracy
3.2.4) Accuracy profile
Accuracy profile was a graphical tool, that shown the confidence interval of the analytical method. It was obtained by linking on one hand the upper Î² acceptation tolerance and on the other hand, the lower Î² acceptation tolerance calculated at each concentration levels for the validation standards (Figure 3). The accuracy profiles for pure water, Seine river water, biofilm and cyanobacteria were constructed and were shown on Figure 3. Analytical procedure is valid as long as the confidence interval (in dashed lines) is included between the both dotted lined, that symbolized the acceptance limits (± 20%). The lower limit of the analytical procedure was given by the intersection between the accuracy profile and the lower acceptance limit (- 20%). Therefore the lower limit of the analytical procedure was 2 ng.ml-1 in pure water and in Seine river water and 0.25 ng.mg-1 dry weight in biofilm and 0.6 ng.mg-1 in cyanobacteria. The accuracy profile in the four matrices tested didn't cross the upper limit of acceptance (+ 20%) in our range of concentration. Therefore the upper limits were chosen to be the highest concentration tested for each matrix ie 13.5 ng.ml-1 for pure and Seine river water, 1.6 ng.mg-1 for biofilm and 4 ng.mg-1 for cyanobacteria.
Figure 1: ESI/MS-MS fragmentation spectra of BMAA, DAB and D3-DAB.
Figure 2: LC/MS-MS chromatogram with specific MRM transitions for BMAA, DAB and D3-DAB in pure water spiked at 25 ng.ml-1 (left column) and in extract of cyanobacteria PCC 6506 after SPE cleaning-up protocol spiked at the same concentration (right column).
Cyanobacteria PCC 6506
Treatment with TCA
(reference in pure water)
Seine river water
- 30 %
- 35 %
Table 2: Precision (repeatability and intermediate precision) obtained in pure water for validation samples using different models of regression for calibration curves for each concentration level tested.
Concentration levels (ng.ml-1)
Models and parameters of regression
y = 7.24E-11 x2 + 1.04E-3 x
r2= 0.9994, RSD = 15.6%
y= 10.772E-3 x -1.4712
r2=0.9984, RSD= 17.2%
Linear through zero (m=4)
y= 9.682E-4 x
r2=0.9988, RSD= 11.0%
Linear through zero (m=2)
y= 10.03E-3 x
r2=0.9997, RSD= 9.9%
IP = Intermediate precision, Repeat = repeatability, p = number of days (series) of analysis, n = number of repetition per day of analysis, m = number of concentration levels used for calibration curves.
Table 3: Details results of validation by total error approach for all matrixes (pure water, Seine river water, biofilm and cyanobacteria)
Mean concentration in ng.ml-1
or in ng.mg-1
Seine River water
Figure 3: Accuracy profile from the validation using linear regression mode in A) pure water, B) Seine river water, C) biofilm and D) PCC 6506. The dots represented the relative back calculated concentration of the validation samples, the plain lines are the relative bias, the dashed lines are the Î² acceptance limits and the dotted curves represent the acceptance limits.
 S.A. Cohen, Analyst, 137 (2012) 1991-2005.
 S. Cadel-Six, C. Peyraud-Thomas, L. Brient, N.T. de Marsac, R. Rippka, A. Méjean, Appl. Environ. Microbiol., 73 (2007) 7605-7614.
 T. Kubo, N. Kato, K. Hosoya, K. Kaya, Toxicon, 51 (2008) 1264-1268.
 A. Li, H. Fan, F. Ma, P. McCarron, K. Thomas, X. Tang, M.A. Quilliam, Analyst, 137 (2012) 1210-1219.
 T. Krüger, B. Mönch, S. Oppenhäuser, B. Luckas, Toxicon, 55 (2010) 547-557.
 E.J. Faassen, F. Gillissen, M. Lürling, PLoS ONE, 7 (2012) e36667.
 P. Hubert, J.-J. Nguyen-Huu, B. Boulanger, E. Chapuzet, N. Cohen, P.-A. Compagnon, W. Dewé, M. Feinberg, M. Laurentie, N. Mercier, G. Muzard, L. Valat, E. Rozet, J Pharm Biomed Anal, 48 (2008) 760-771.
 C.T. Viswanathan, S. Bansal, B. Booth, A.J. DeStefano, M.J. Rose, J. Sailstad, V.P. Shah, J.P. Skelly, P.G. Swann, R. Weiner, Pharm. Res., 24 (2007) 1962-1973.