Screening Method for 23 Alkaloids in Human Serum Using LC/MS/MS with a Pentafluorophenyl Column in Dynamic Multiple Reaction Monitoring Mode

Alkaloids are nitrogen-containing organic compounds, generally basic, and found in plants, fungi, and bacteria. Some alkaloids are used in medicine, but some compounds are highly toxic. Accidental ingestion, homicide, and suicide have occurred due to plants containing alkaloids. The identification of toxic components in biological samples is important for the diagnosis and/or treatment of poisoning cases in forensic and emergency medicine. Alkaloids have a wide variety of structures, such as isoquinoline alkaloid, indole alkaloid, tropane alkaloid, and diterpene alkaloid; therefore, there are few reports of simultaneous analysis methods. We have established a method for the simultaneous analysis of 23 alkaloids in human serum with a liquid chromato-graph-tandem mass spectrometer (LC/MS/MS). A liquid-liquid extraction which was modified from the first step of the QuEChERS AOAC method was used for serum pretreatment. The separation of the compounds was performed using a pentafluorophenyl (PFP) column, CAPCELL CORE PFP (2.1 mm I.D. × 100 mm, 2.7 μm) in gradient mode. Mobile phase A consisted of 10 mM ammonium formate and 0.1% formic acid in ultrapure water, and mobile phase B was 10 mM ammonium formate and 0.1% formic


Introduction
Alkaloids are nitrogen-containing organic compounds, which are generally basic, and found in plants, fungi, and bacteria [1]. Although often used as medicines, the use of some alkaloids has been reported in homicide and suicide cases owing to their strong toxicity [2] [3] [4] [5] [6]. In addition, there have been many cases of alkaloid poisoning caused by ingesting plants containing toxic alkaloids that were mistaken as edible [7]- [22]. In poisoning instances, it is important to identify the causative compounds to diagnose toxic symptoms and/or determine the cause of death. Screening tests for toxicants in biological samples are required, particularly where the identity of the ingested plant is unknown.
Therefore, we developed a screening method for alkaloids contained in plants posing the risk of accidental ingestion. There are various types of alkaloids, such as isoquinoline alkaloid, indole alkaloid, tropane alkaloid, and diterpene alkaloid. Many reports of analytical methods exist for detecting individual alkaloids [3] [4] [21]- [31], but there have been few reports of their simultaneous analysis [32] because of their numerous chemical structures. Simultaneous analysis of alkaloids in human urine and herbal samples by liquid chromatograph-tandem mass spectrometer (LC/MS/MS) using a C8 column was reported in one study [32]. In forensic medicine, the results of toxicology tests in blood or serum are important for diagnosing poisoning symptoms at the time of death, so we developed an analysis method using serum as the specimen. The analytical column commonly used for medicinal and toxic substance analysis is alkyl-silica columns; however, peak tailing occurs with some basic compounds, including alkaloids. Therefore, we validated an analytical method using a pentafluorophenyl (PFP) analytical column, which efficiently separated polar aromatic compounds. In recent years, PFP columns have been applied to various analyses [33] [34] [35] [36] [37], as they exhibit different retention behavior from columns of alkyl phases, which retain polar aromatic compounds, demonstrate excellent stereoselectivity, and effectively separate halogen compounds.
The target compounds of the present method were 23 alkaloids found in plants. Table 1 shows the classification of the alkaloids, families, and representative plants for each compound. The measurement mode employed was dynamic multiple reaction monitoring (dMRM). The dMRM mode only operates when the target compounds are eluted from liquid chromatograph (LC) and are more sensitive than the multiple reaction monitoring (MRM) mode. The serum was pretreated using a modified QuEChERS AOAC method [38]. We established a screening method for 23 alkaloids in human serum using dMRM with LC/MS/MS Pouches, AOAC method (containing magnesium sulfate (6.0 g) and sodium acetate (1.5 g)) were purchased from Agilent Technologies (Santa Clara, CA, USA). L-Consera ® I EX Nissui (Nissui Pharmaceutical Co., Ltd., Tokyo, Japan), a freeze pool serum, was used as the blank serum. The blank serum was confirmed to be negative for the analytes of interest before use.

Preparation of Standard Solutions and Quality Control Samples
Aconitum diester alkaloids standard and aconitum monoester alkaloids standard were dissolved in 5 mL of acetonitrile and stored at -20˚C prior to use. Stock standard solutions of 500 µg/mL of reserpine were prepared in acetonitrile and stored at -20˚C prior to use. Stock solutions of other compounds (500 µg/mL) were prepared by dissolving the corresponding standards in methanol and stored at -20˚C prior to use. The internal standard (IS), galantamine-d 6 , was dissolved in methanol to prepare a 1.0 mg/mL stock solution, diluted to 200 ng/mL with acetonitrile, and stored at -20˚C prior to use. To prepare the quality control samples, each stock solution was diluted to 60, 600, and 6000 ng/mL for hypaconitine; 40, 400, and 4000 ng/mL for mesaconitine; 38, 378, and 3782 ng/mL for 14-anisoylaconine; 19, 188, and 1880 ng/mL for benzoylhypaconine; 38, 377, and 3767 ng/mL for benzoylmesaconine; and 20, 200, and 2000 ng/mL for the other compounds with acetonitrile at the time of use, and an aliquot of each diluted solution (10 µL) was added to the blank serum (0.2 mL).

Sample Preparation
The samples were prepared using a modified version of the procedure described by Kudo et al. [38]. Serum (0.

Method Validation
The matrix effect, extraction recovery, and lower limit of detection (LOD) were

Matrix effect (%) = (B/A) × 100
Recovery rate (%) = (C/B) × 100 The LOD was defined as the lowest concentration that yielded a reproducible peak with a signal-to-noise ratio of 3 or more.

Application to Forensic Case
The validated screening method was used to analyze a sample that had been previously qualitatively tested using another method. The serum used was residual samples that had been stored at -30˚C. Case: The deceased was a female in her late teens. She was found deceased, lying on her back, on the floor in her home. The vomit on the floor was mixed with leaves and roots shaped like aconite. It was assumed that the deceased ate aconite for suicidal purposes. The deceased's specimen was analyzed using another method dedicated to aconite at the time of appraisal in order to investigate the cause of death, and aconitine and jesaconitine was detected.

LC/MS/MS Conditions
The optimized MS/MS parameters for the compounds are shown in Table 2. Figure 1 shows the dMRM chromatograms of the serum extracts prepared by adding the compounds to the blank serum at a concentration of 150 ng/mL for hypaconitine, 100 ng/mL for mesaconitine, 95 ng/mL for 14-anisoylaconine, 47 ng/mL for benzoylhypaconine, 94 ng/mL for benzoylmesaconine, and 50 ng/mL for other compounds using the pretreatment procedure. Satisfactory separation of all compounds was observed with no interference between the peaks in the dMRM chromatograms of the serum extracts. The CAPCELL CORE PFP column is a reverse-phase core shell column in which a PFP group is added to the silica gel. PFP columns allow for dipole-dipole, π-π, charge transfer, and ionexchange interactions, in addition to the dispersive interactions available on alkyl-silica columns [39]. When methanol was used as the mobile phase organic solvent, the peak areas were larger than those when acetonitrile was used. It was speculated that the π-π interactions functioned by using methanol. Hence, we employed methanol as the organic solvent for mobile phase B. Table 3 shows the retention time, matrix effect, extraction recovery, and LOD of the compounds. The matrix effects of aconitum diester alkaloids including aconitine, hypaconitine, jesaconitine, and mesaconitine were 53.1% -65.2%. The retention times of these compounds were longer than 11 min, and it was estimated that ion suppression occurred due to overlap in retention time with American Journal of Analytical Chemistry   Table 3. Retention times, matrix effect (n = 6), extraction recovery (n = 6), and lower limit of detection (LOD) for analyte in human serum. to its susceptibility to oxidation [42]. The recovery rates of reserpine were 40.0%

Sample Preparation, Matrix Effect, Extraction Recovery, and LOD
-54.5%, and it was thought to be photodegraded during extraction because the compounds decompose under light [43]. The recovery rates of other compounds were 69.1% -111.6%, confirming that the target substances were sufficiently ex-American Journal of Analytical Chemistry tracted. Aconitum diester alkaloids, papaverine, and reserpine were detected with an LOD lower than their intoxication ranges [44] [45]. Therefore, the present method was determined to be compatible with the screening analysis for 23 alkaloids in the serum.
The QuEChERS method is a pretreatment procedure used to extract pesticide residues, such as those found in vegetables [46] [47]. Kudo et al. applied the liquid-liquid extraction method, the first step of the QuEChERS method, in the analysis of human biological samples, including blood [38]. Since aconitines are degraded by methanol [48], we modified the method to use acetonitrile as the redissolving solution. This simple method which is performed using the dehydration and the salting-out effects of magnesium sulfate and sodium acetate, respectively, was suitable for the simultaneous extraction of compounds with diverse chemical structures. In dMRM mode, analytes are only monitored while they are eluting from the LC, and the MS duty cycle is not wasted by monitoring for the analytes when they are not expected. Hence, the peak areas of the compounds using dMRM are larger than those obtained using MRM. Thus, the present method was able to detect the compounds at an LOD of 0.01 -0.6 ng/mL (Table 3).
By using the present pretreatment method and LC/MS/MS analysis with the PFP analytical column, we succeeded in simultaneously analyzing 23 alkaloids that had previously been analyzed using each different method.

Application to Forensic Cases
The serum of the deceased, who was presumed to have ingested aconite, was analyzed using the validated method. The dMRM chromatogram of the extract obtained from the serum of the deceased is shown in Figure 2. Aconitine and jesaconitine were detected, with the result being the same as that of another analysis method for aconite performed during autopsy. The present method was successfully applied to the LC/MS/MS screening of serum for forensic cases. As the compounds and composition of the aconite toxin vary by species and region [49], several compounds must be included in the screening analysis to avoid missing the causative agent of poisoning. The present method was found to be effective for alkaloid screening in human serum.

Conclusion
In this study, we established an LC/MS/MS simultaneous analysis method for 23 alkaloids in human serum, which were difficult to analyze simultaneously. The pretreatment method was a modification of the first step of the QuEChERS AOAC method. Twenty-three alkaloids were efficiently separated by using a PFP analytical column. Simultaneous analysis using the dMRM mode improved the sensitivity. The validated method was applied to a forensic case and successfully screened alkaloids in human serum. Hence, the present LC/MS/MS simultaneous analysis method was effective for the screening of alkaloids in human serum in forensic and clinical toxicology.

Ethical Approval
This study has been approved by the Hospital Ethics Committee of the Kitasato University School of Medicine (B20-082).