Development and Optimization of a Simple and Sensitive Method for the Determination of Virginiamycin M1 Antibiotic in Aqueous Media by Capillary Electrophoresis

Virginiamycin antibiotic is used in Mexico for animal growth promotion, but has been banned in Europe due to risk of resistance development. Giving that monitoring of antibiotics is critical, the objective of this work was to develop an analytical method based on capillary electrophoresis and liquid-liquid extraction, to quantify virginiamycin antibiotic in livestock wastes. Linearity, precision, bias and extraction parameters were evaluated following ISO/IEC 17025 procedures. Method and extraction validation was satisfactorily, with average values of abso-lute recovery AR = 86% , extraction e ﬃ ciency EE = 87% , and matrix e ﬀ ects ME = -14% , RSD < 15% for concentrations as low as 50 µ g/L.


Introduction
Virginiamycin, an antibiotic belonging to streptogramin class, and mixture of two components: factors M1 and S1, acts individually against Micrococcus aureus and Bacillus subtilis, but in mixtures becomes highly effective against vancomycin and methicillin-resistant infections in humans, especially from Enterococcus faecium and Staphylococcus aureus (Hammerum et al, 1998).
Relevant properties and structure for M1 are depicted in Figure 1. Composed of lactone and containing pyrrolidine and oxazole rings and equal to mikamycin A, ostreogrycin A, streptogramin A and pristinamycin IIA (Goffic et al, 1977), M1 factor assembles a lipophilic, neutral, and low-degradable molecule mainly evacuated in excreta without relevant assimilation in animal tissues.
Mexico and the USA have approved virginiamycin as a feedstuff supplement, alone or combined with other agents in cattle, pig and poultry for disease prevention, increased weight growth and fodder efficiency, at doses ranging from 5 g/ton to 25 g/ton of food (Anadón and Martínez-Larrañaga, 1999).
These growth promoters are administered for prolonged periods of time at sub-therapeutic concentrations. However, with such scheme arises an increased risk of inducing antimicrobial resistance in animal pathogens, which could colonize humans and transfer their resistance to other microorganisms (Wegener, 2003). Evidence of adverse human health consequences include infections that probably would not occur, as with fluoroquinolone-resistant Salmonella cases (Ventola, 2015). Several researchers argue that resistant infections in clinical settings have been due to migration of genetic determinants, between environmental bacteria and human pathogens (Martinez, 2009b;Wright, 2010).
Due to incorrect use of veterinary antibiotics, resistant bacteria populate the environment, animals and food products, increasing pathogen exposure and triggering human infection difficult to treat (Martinez, 2009a).
Europe banned use of virginiamycin for animal growth in 1998, due to concerns about impairing the efficacy of streptogramin to treat human infections due to rise of resistance (Aarestrup et al, 2001). However, Mexico and other countries consider virginiamycin safe and authorize it as a growth promoter.
Because most of virginiamycin intake is excreted to environment (de Sanidad Inocuidad y Calidad Agroalimentaria, 2016), the risk of resistance development in environmental bacteria arise, justifying an antibiotic monitoring as part of a surveillance program (Sandegren, 2014;Bellanger et al, 2014;Bengtsson-Palme and Larsson, 2015). In fact, different studies revealed selection of resistant bacteria at low concentrations, like those found in aquatic environments receiving agricultural effluents (Hernando et al, 2006;Martinez, 2009b).

Development of Analytical Method
To determine the most significant factors of running buffer that affects analyte separation, an experimental design approach was chosen, as it enables lower time consumption and best data analysis. A screening method based on a 2 4 1 reduced factorial design was set-up to identify the most significant buffer variables at two levels: A-pH; B-SDS; C-phosphate and D-β-cyclodextrin, as shown in Table 1. Response variables H and N were computed using Design Expert 7.0 software. Other variables, were optimized modifying one variable at a time with a constant running buffer. The selected variables were: A-NaCl percent added to standard; B-separation voltage; C-injection time and Dcartridge temperature. Selection of the best conditions was done comparing electropherograms and peak efficiencies.

CE Method Validation
The CE method was validated by evaluating precision, linear range, limit of detection (LOD), limit of quantitation (LOQ) and bias. An Analysis of Variance (ANOVA) was used to compare variation within groups at 95 percent confidence. The LOD is the lowest concentration that could be detected with 277  278  279  280  281  282  283  284  285  286  287  288  289  290  291  292  293  294  295  296  297  298  299  300  301  302  303  304  305  306  307  308  309  310  311  312  313  314  315  316  317  318  319  320  321  322 Springer Nature 2021 L A T E X template Research Article 7 enough confidence; it was approximated by diluting a calibrator solution until signal of analyte is similar to noise, then analyzed ten times and calculated as LOD =3⇥ S 0 o ,w h e r eS 0 o = S p n explains data disperssion. The limit of quantification (LOQ) was then estimated as LOD ⇥ 5.
To quantify virginiamycin, a linear regression of measured absorbance against concentration was built using seven calibrators determined by triplicates at analyte concentrations of 0.05, 0.1, 0.2, 0.4, 0.8, 1.6 and 3.2 mg/L.
Precision was evaluated analyzing single and intra-day repeatability of the smaller calibrator along with two midpoints, running five and three repetitions respectively. For inter-day repeatability, analysis was done at three days at constant conditions. One-factor ANOVA was run to test for variability.
Repeatability limit (r) was calculated as r =2 .8 ⇥ S r .w h e r eS r is the standard deviation of repeatability. Intermediate precision was obtained by the square root of the sum of the squares of within-group and between-group precision. Then standard deviation of the repeatability S r , was obtained by calculating the square root of the within-group mean square term, which represents the within-group variance as S r = p MS i . Contribution to the total variance of the clustering factor S i is also obtained from the ANOVA test: . The intermediate precision S I was calculated by combining both terms: S I = p S 2 r + S 2 i . Finally, bias was measured in relative terms as a percentage of recovery R'= x' x x add ⇥ 100 using blank and spiked samples, where x' is the average spiked value, x is te average value with the added concentration and x add is the added standard (Magnusson, 2014).
Such procedure is based on the formation of an aqueous two-phase system To form the two phases, the substances in the vials were dissolved aided by a sonication bath for 5 min. Then, vials were centrifuged for 3 minutes and an aliquot of 10 µL was transferred to a microvial. Finally, 90 µL of diluent was added, vorterized and ready for injection in the CE system. Fortified samples were prepared in the same fashion, but adding antibiotics to diluent. Performance was verified using the parameters of extraction efficiency EE, matrix effects ME and recovery AR, as disclosed in following equations: 3 Results

Method development
Initially, the MEKC method was initiated with the injection of a standard virginiamicyn M1 solution dissolved in distilled water. A typical electropherogram for the separation is presented in Figure 3 at a concentration of 20 mg/L, at t m around 13 min.
Then, a factorial design was employed to determine which BGB conditions affects H and N. The results are presented in Table 2  Goodness of fit test also suggests adequate model adjustment and precision, with Adj.R 2 > 0.8 and RSD < 15%. Regarding optimization of NaCl added to sample and injection volume, addition of NaCl from 0.1% to 1% to standard solution increased N massively; as for the standard peak N rises up to 7 ⇥ 10 5 .
In this MEKC method, online concentration was based on salt addition to samples, in a manner ionic strength scale up two to three times higher than BGB (Palmer et al, 1999).

Method Validation
Validation parameters evaluated were precision, working range, bias, LOD and LOQ. Linear regression and goodness of fit are presented in Table 4. Working range was validated from 0.05 mg/L to 3.2 mg/L. Repeatability and intermediate precision were 9.57% and 11.57% RSD respectivelty along the working range. Bias was calculated as R%= 8.51 and RSD% = 12.16, which represent acceptable values. LOD ⇡ 0.093 mg/L and LOQ ⇡ 0.047mg/L were also estimated. Figure 5 shows an electropherogram for blank and spiked sample at 5 mg/L.

Extraction Validation
As noted, not major signal suppression appears at 1:9 weight ratio dilution, though spiked samples have to be analyzed for every run. Further, resolution R s 1.5 seems reasonable to detect and quantify virginiamycin M1. To verify the identity of analyte, Figure 6 shows EG's of samples fortified before and after extraction, at 50 µg/L and 500 µg/L respectively. Virginiamycin peak appears labeled and their vicinity is spanned in the bottom EG, showing the identity of virginiamycin unambiguously. Extraction data is presented in Table 5 for three concentrations. The computed parameters were: EE =0.87, ME = 0.14 and AR =0 .86. Since signal suppression stand low, no significant matrix effects reduce analyte signal. Also, average values of EE and AR found satisfactory, allowing most analyte recovered from extraction procedure.

Consent to Participate
Not applicable.

Consent to Publish
Not applicable.

Availability of data and materials
The data that support the findings of this study are available from the corresponding author C.Díaz-Quiroz, upon reasonable request.

Authors Contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Juan Francisco Hernández-Chávez and Jesús Fernando Robles-Castro. The first draft of the manuscript was written by Carlos Abraham Díaz-Quiroz. Gabriela Ulloa-Mercado and Carlos Díaz contributed to original concept, funding, supplied reagent and manuscript revisions and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.