*Article* **Simultaneous Determination of Caffeine and Paracetamol in Commercial Formulations Using Greener Normal-Phase and Reversed-Phase HPTLC Methods: A Contrast of Validation Parameters**

**Prawez Alam 1,\* , Faiyaz Shakeel 2 , Abuzer Ali 3 , Mohammed H. Alqarni 1 , Ahmed I. Foudah 1 , Tariq M. Aljarba 1 , Faisal K. Alkholifi 4 , Sultan Alshehri 2 , Mohammed M. Ghoneim <sup>5</sup> and Amena Ali 6**


**Abstract:** There has been no assessment of the greenness of the described analytical techniques for the simultaneous determination (SMD) of caffeine and paracetamol. As a result, in comparison to the greener normal-phase high-performance thin-layer chromatography (HPTLC) technique, this research was conducted to develop a rapid, sensitive, and greener reversed-phase HPTLC approach for the SMD of caffeine and paracetamol in commercial formulations. The greenness of both techniques was calculated using the AGREE method. For the SMD of caffeine and paracetamol, the greener normal-phase and reversed-phase HPTLC methods were linear in the 50–500 ng/band and 25–800 ng/band ranges, respectively. For the SMD of caffeine and paracetamol, the greener reversed-phase HPTLC approach was more sensitive, accurate, precise, and robust than the greener normal-phase HPTLC technique. For the SMD of caffeine paracetamol in commercial PANEXT and SAFEXT tablets, the greener reversed-phase HPTLC technique was superior to the greener normal-phase HPTLC approach. The AGREE scores for the greener normal-phase and reversedphase HPTLC approaches were estimated as 0.81 and 0.83, respectively, indicated excellent greenness profiles for both analytical approaches. The greener reversed-phase HPTLC approach is judged superior to the greener normal-phase HPTLC approach based on numerous validation parameters and pharmaceutical assays.

**Keywords:** caffeine; greener HPTLC; paracetamol; simultaneous determination; validation

#### **1. Introduction**

Paracetamol (Figure 1A) is the commonly administered anti-inflammatory and antipyretic medicine, especially in case of pediatric and geriatric patients [1,2]. It is commercially available in a wide range of dosage forms [2]. Caffeine (Figure 1B) is a pseudoalkaloidal drug that is commonly used in combination with paracetamol [3,4]. The combination of paracetamol and caffeine is the world's most widely used combination [4]. As

**Citation:** Alam, P.; Shakeel, F.; Ali, A.; Alqarni, M.H.; Foudah, A.I.; Aljarba, T.M.; Alkholifi, F.K.; Alshehri, S.; Ghoneim, M.M.; Ali, A. Simultaneous Determination of Caffeine and Paracetamol in Commercial Formulations Using Greener Normal-Phase and Reversed-Phase HPTLC Methods: A Contrast of Validation Parameters. *Molecules* **2022**, *27*, 405. https:// doi.org/10.3390/molecules27020405

Academic Editors: Victoria Samanidou and Natasa Kalogiouri

Received: 16 December 2021 Accepted: 6 January 2022 Published: 9 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

a result, the qualitative and quantitative standardization of caffeine and paracetamol in commercially available formulations is necessary. 

‐

**Figure 1.** Chemical structures of (**A**) paracetamol and (**B**) caffeine.

‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ An extensive literature search revealed various analytical approaches for the simultaneous determination (SMD) of caffeine and paracetamol in commercial formulations and biological fluids. For the SMD of caffeine and paracetamol in commercial formulations, different spectrometry techniques involving various chemical procedures, such as derivatization have been used [5–10]. For the SMD of caffeine and paracetamol in various commercial dosage forms, several high-performance liquid chromatography (HPLC) techniques have been used [4,11–19]. Caffeine and paracetamol were also quantified simultaneously in a human plasma sample using a HPLC method [19]. For the SMD of caffeine and paracetamol in human plasma samples, a liquid-chromatography mass-spectrometry (LC–MS) technique was also used [20]. For the SMD of caffeine and paracetamol in their pure forms and formulations, certain high-performance thin-layer chromatography (HPTLC) techniques have been used [21–23]. Various voltammetry-based approaches have also been applied for the SMD of caffeine and paracetamol in their dosage forms [24–27]. Dual-mode gradient HPLC and HPTLC methods have also been used for the SMD of caffeine and paracetamol in the presence of paracetamol impurities [28]. The electrospray laser desorption ionization mass spectrometry technique was also utilized for the SMD of caffeine and paracetamol in tablets [29]. An electrochemical cell-on-a-chip device fabricated using 3D-printing technology was also used for the SMD of caffeine and paracetamol [30]. A genetic algorithm based on wavelength selection was also applied for the SMD of caffeine and paracetamol [31]. Some other approaches, such as near-infrared spectrometry [32], flow-injection spectrometry [33], micellar liquid chromatography [34], and micellar electrokinetic capillary chromatography [35] approaches were also proposed for the SMD of caffeine and paracetamol in their dosage forms. Published reports on the SMD of caffeine and paracetamol suggested various analytical approaches for their analysis. However, the greenness scale of any of the reported analytical approach was not estimated. In addition, greener HPTLC approaches have not been utilized for the SMD of caffeine and paracetamol. For the estimation of the greenness scale, different quantitative analytical methodologies have been presented [36–40]. For the estimation of the greenness scale, only the "Analytical Greenness (AGREE)" analytical approach considers all twelve green analytical chemistry (GAC) principles [38]. As a result, the AGREE analytical methodology was applied for the estimation of greenness scale of the greener normal-phase and reversed-phase HPTLC approaches [38].

‐ In comparison to the greener normal-phase HPTLC approach, the current study intends to establish and validate a rapid, sensitive, and greener reversed-phase HPTLC approach for the SMD of caffeine and paracetamol in commercial formulations. Following "The International Council for Harmonization (ICH)" Q2-R1 recommendations, the greener normal-phase and reversed-phase HPTLC methods for the SMD of caffeine and paracetamol were validated [41].

‐ ‐

‐

‐ ‐

‐

#### **2. Results and Discussion**

‐ ‐

#### *2.1. Method Development*

For the development of a suitable band for the SMD of caffeine and paracetamol using the greener normal-phase HPTLC technique, various amounts of ethyl acetate (EA) and ethanol (E), including EA/E (50:50, *v*/*v*), EA/E (60:40, *v*/*v*), EA/E (70:30, *v*/*v*), EA/E (80:20, *v*/*v*), EA/E (85:15, *v*/*v*), and EA/E (90:10, *v*/*v*) were studied as the greener mobile phases. ‐ 

The greener mobile phases, such as EA/E (50:50, *v*/*v*), EA/E (60:40, *v*/*v*), EA/E (70:30, *v*/*v*), EA/E (80:20, *v*/*v*), and EA/A (90:10, *v*/*v*) revealed poor chromatographic peaks of caffeine and paracetamol with high asymmetry factor (As) for caffeine (As > 1.25) and paracetamol (As > 1.30). When the greener mobile phase EA/E (85:15, *v*/*v*) was evaluated, it was discovered that this greener mobile phase provided well-resolved and intact chromatographic peaks for caffeine at a retardation factor of (R<sup>f</sup> ) = 0.40 ± 0.01 and for paracetamol of R<sup>f</sup> = 0.59 ± 0.02 (Figure 2). Caffeine and paracetamol were also predicted to have As values of 1.06 and 1.08, respectively, which are very trustworthy. As a consequence, the EA/E (85:15, *v*/*v*) was chosen as the final mobile phase for the SMD of caffeine and paracetamol in commercial tablets utilizing the greener normal-phase HPTLC method. ‐ ‐ ‐ ‐ ‐ ‐

‐ ‐ ‐ **Figure 2.** Normal-phase high-performance thin-layer chromatography (HPTLC) chromatogram of standard caffeine and paracetamol.

‐ ‐ For the development of a suitable band for the SMD of caffeine and paracetamol using the greener reversed-phase HPTLC technique, various amounts of E and water (W), including E/W (50:50, *v*/*v*), E/W (60:40, *v*/*v*), E/W (70:30, *v*/*v*), E/W (80:20, *v*/*v*), and E/W (90:10, *v*/*v*) were studied as the greener mobile phases. All of the green mobile phases investigated were created under chamber saturation conditions (Figure 3).

‐ ‐ The greener mobile phases, such as E/W (60:40, *v*/*v*), E/W (70:30, *v*/*v*), E/W (80:20, *v*/*v*), and E/W (90:10, *v*/*v*) revealed poor chromatographic peaks of caffeine and paracetamol with poor As for caffeine (As > 1.30) and paracetamol (As > 1.35). When the greener mobile phase E/W (50:50, *v*/*v*) was evaluated, it was discovered that this greener mobile phase provided well-resolved and intact chromatographic peaks of caffeine at R<sup>f</sup> = 0.43 ± 0.01 and of paracetamol at R<sup>f</sup> = 0.57 ± 0.02 (Figure 4). Caffeine and paracetamol were also predicted to have As values of 1.10 and 1.09, respectively, which are very trustworthy. As a consequence, the E/W (50:50, *v*/*v*) was chosen as the final mobile phase for the SMD of caffeine and paracetamol in commercial tablets utilizing the greener reversed-phase HPTLC method. The maximum response was obtained at a wavelength of 260 nm for caffeine and paracetamol when the spectral bands for caffeine and paracetamol

were recorded using densitometry mode. As a result, the whole SMD of caffeine and paracetamol took place at 260 nm. ‐

‐

‐

‐

‐ ‐ ‐ **Figure 3.** Developed thin-layer chromatography (TLC) plate for standard caffeine, standard paracetamol, commercial tablets PANEXT, and commercial tablets SAFEXT developed using ethanol (E)/water (W) (50:50 *v*/*v*) as the greener mobile phase for the greener reversed-phase HPTLC method. ‐

#### ‐ *2.2. Method Validation*

‐ ‐ ‐ ‐ ‐ ‐ The ICH-Q2-R1 guidelines were used to estimate various parameters for the SMD of caffeine and paracetamol [41]. The results of the linear regression analysis of caffeine and paracetamol calibration curves utilizing the greener normal-phase HPTLC technique are summarized in Table 1. Caffeine and paracetamol calibration curves were linear in the 50–500 ng/band range for both drugs. Caffeine and paracetamol's determination coefficients (R<sup>2</sup> ) were found to be 0.9928 and 0.9970, respectively. Caffeine and paracetamol's regression coefficients (R) were found to be 0.9963 and 0.9984, respectively. The values of R<sup>2</sup> and R were highly significant for both the compounds (*p* < 0.05). These findings suggested a strong link between the concentration and measured response of caffeine and paracetamol. All these findings indicated the reliability of the greener normal-phase HPTLC approach for the SMD of caffeine and paracetamol.


**Table 1.** Results of the linear regression analysis for the simultaneous determination (SMD) of caffeine and paracetamol using the greener normal-phase high-performance thin-layer chromatography (HPTLC) method (mean ± SD; *n* = 6).

R 2 : determination coefficient; R: regression coefficient; LOD: limit of detection; LOQ: limit of quantification.

The resulting data for the linear regression analysis of caffeine and paracetamol calibration curves utilizing the greener reversed-phase HPTLC technique are summarized in Table 2. Caffeine and paracetamol calibration curves were linear in the 25–800 ng/band range for both drugs. Caffeine and paracetamol's R<sup>2</sup> were found to be 0.9976 and 0.9966, respectively. Caffeine and paracetamol's R were found to be 0.9987 and 0.9982, respectively. The values of R<sup>2</sup> and R were highly significant for both the compounds (*p* < 0.05). These findings again suggested a strong link between the concentration and measured response of caffeine and paracetamol. All these findings indicated the reliability of the greener reversed-phase HPTLC technique for the SMD of caffeine and paracetamol. However, the greener reversed-phase HPTLC technique was more linear than the greener normal-phase HPTLC technique.

**Table 2.** Results for linear regression analysis for the SMD of caffeine and paracetamol using the greener reversed-phase HPTLC method (mean ± SD; *n* = 6).


R 2 : determination coefficient; R: regression coefficient; LOD: limit of detection; LOQ: limit of quantification.

The parameters of the system appropriateness for the greener normal-phase HPTLC methodology are summarized in Table 3. For the SMD of caffeine and paracetamol, the R<sup>f</sup> , As, and number of theoretical plates per meter (N/m) for the greener normal-phase HPTLC technique were determined to be satisfactory. The parameters of the system appropriateness for the greener reversed-phase HPTLC methodology are summarized in Table 4. For the SMD of caffeine and paracetamol, the R<sup>f</sup> , As, and N/m for the greener reversed-phase HPTLC technique were also determined to be satisfactory.

**Table 3.** System suitability parameters in terms of retardation factor (R<sup>f</sup> ), asymmetry factor (As), and a number of theoretical plates per meter (N/m) of caffeine and paracetamol for the greener normal-phase HPTLC method (mean ± SD; *n* = 3).


**Table 4.** The R<sup>f</sup> , As, and N/m values of caffeine and paracetamol for the greener reversed-phase HPTLC method (mean ± SD; *n* = 3).


For assessing caffeine and paracetamol, the percent of recovery was utilized to estimate the accuracy of the greener normal-phase and reversed-phase HPTLC techniques. The accuracy evaluation results for the greener normal-phase HPTLC technique are summarized in Table 5. Using the greener normal-phase HPTLC technique, the percent recoveries of caffeine and paracetamol at three separate quality control (QC) samples were expected to be 97.13–104.88 and 96.57–103.23 percent, respectively. The accuracy evaluation results for the greener reversed-phase HPTLC technique are summarized in Table 6. Using the greener reversed-phase HPTLC technique, the percent recoveries of caffeine and paracetamol at three separate QC samples were expected to be 98.84–100.62 and 98.60–101.50 percent, respectively. These results showed that both analytical techniques were accurate for the SMD of caffeine and paracetamol. For the SMD of caffeine and paracetamol, however, the greener reversed-phase HPTLC methodology was more accurate than the greener normal-phase HPTLC methodology.

**Table 5.** Measurement of the accuracy of caffeine and paracetamol for the greener normal-phase HPTLC method (mean ± SD; *n* = 6).


CV: coefficient of variance.

The precision of the greener normal-phase and reversed-phase HPTLC techniques was investigated as intra/inter-assay precision and given as a percent of the coefficient of variation (CV) for the SMD of caffeine and paracetamol. Table 7 summarizes the results of intra/inter-day precisions for the SMD of caffeine and paracetamol using the greener normal-phase HPTLC technique. The percent CVs of caffeine and paracetamol for the intra-day variation were estimated as 1.30–2.39 and 1.91–3.42 percent, respectively. The percent CVs of caffeine and paracetamol for inter-day variation were estimated as 1.51–2.55 and 1.86–3.56 percent, respectively. Table 8 summarizes the results of intra/inter-day precisions for the SMD of caffeine and paracetamol using the greener reversed-phase HPTLC technique. The percent CVs of caffeine and paracetamol for the intra-day variation were estimated as 0.40–0.85 and 0.52–0.96 percent, respectively. The percent CVs of caffeine and paracetamol for inter-day variation were estimated as 0.42–0.78 and 0.55–1.03 percent, respectively. These findings indicated that both the analytical approaches were precise for the SMD of caffeine and paracetamol. However, the greener reversed-phase HPTLC methodology was more precise than the greener normal-phase HPTLC methodology for the SMD of caffeine and paracetamol.

**Table 6.** Measurement of the accuracy of caffeine and paracetamol for the greener reversed-phase HPTLC method (mean ± SD; *n* = 6).


CV: coefficient of variance.

**Table 7.** Assessment of intra/inter-day precision of caffeine and paracetamol for the greener normalphase HPTLC method (mean ± SD; *n* = 6).


CV: coefficient of variance.

**Table 8.** Assessment of intra/inter-day precision of caffeine and paracetamol for the greener reversedphase HPTLC method (mean ± SD; *n* = 6).


CV: coefficient of variance.

By introducing slight deliberate modifications in the greener mobile phase components, the durability of the greener normal-phase and reversed-phase HPTLC techniques for the SMD of caffeine and paracetamol was examined. Table 9 summarizes the results of robustness evaluation using the greener normal-phase HPTLC approach. The percent

CVs for caffeine and paracetamol were estimated as 2.17–3.33 and 2.48–2.64 percent, respectively. Caffeine and paracetamol R<sup>f</sup> values were also estimated to be 0.39–0.41 and 0.58–0.60, respectively.

**Table 9.** Results of robustness analysis of caffeine and paracetamol for the greener normal-phase HPTLC method (mean ± SD; *n* = 6).


CV: coefficient of variance; R<sup>f</sup> : retardation factor.

Table 10 summarizes the results of robustness evaluation utilizing the greener reversedphase HPTLC methodology. The percent CVs for caffeine and paracetamol were estimated as 0.91–0.94 and 0.95–1.04 percent, respectively. Caffeine and paracetamol R<sup>f</sup> values were also estimated to be 0.42–0.44 and 0.56–0.58, respectively. These results showed that both analytical techniques were reliable for the SMD of caffeine and paracetamol. For the SMD of caffeine and paracetamol, however, the greener reversed-phase HPTLC approach was more robust than the greener normal-phase HPTLC approach.

**Table 10.** Results of robustness analysis of caffeine and paracetamol for the greener reversed-phase HPTLC method (mean ± SD; *n* = 6).


CV: coefficient of variance; R<sup>f</sup> : retardation factor.

The "limit of detection (LOD) and limit of quantification (LOQ)" were used to evaluate the sensitivity of the greener normal-phase and reversed-phase HPTLC methods for the SMD of caffeine and paracetamol. The predicted values of "LOD and LOQ" for caffeine and paracetamol utilizing the greener normal-phase HPTLC technique are summarized in Table 1. Using the greener normal-phase HPTLC technique, the "LOD and LOQ" for caffeine were estimated to be 16.84 ± 0.27 and 50.52 ± 0.81 ng/band, respectively. Using the greener normal-phase HPTLC technique, the "LOD and LOQ" for paracetamol were estimated to be 17.05 ± 0.31 and 51.15 ± 0.93 ng/band, respectively. The predicted values of "LOD and LOQ" for caffeine and paracetamol utilizing the greener reversed-phase HPTLC technique are summarized in Table 2. Utilizing the reversed-phase HPTLC technique, the "LOD and LOQ" for caffeine were estimated to be 8.52 ± 0.12 and 25.56 ± 0.36 ng/band, respectively. Using the greener reversed-phase HPTLC technique, the "LOD and LOQ" for paracetamol were estimated to be 8.71 ± 0.13 and 26.13 ± 0.39 ng/band, respectively. These data suggested that both analytical techniques were sensitive enough for the SMD of caffeine and paracetamol. For the SMD of caffeine and paracetamol, however, the reversed-phase HPTLC methodology was more sensitive than the normal-phase HPTLC methodology. ‐ ‐ ‐

‐ ‐

‐ ‐

‐ ‐

‐ ‐

‐ ‐

‐

−

−

‐

‐

‐

‐

By comparing the R<sup>f</sup> values and superimposed ultra-violet (UV)-absorption spectra of caffeine and paracetamol in the commercial tablets PANEXT and SAFEXT with that of standards caffeine and paracetamol, the specificity of the greener HPTLC approach for the SMD of caffeine and paracetamol was assessed. The overlaid UV spectra of standards caffeine and paracetamol, as well as caffeine and paracetamol in the commercial tablets PANEXT and SAFEXT, are shown in Figure 5. ‐ ‐

**Figure 5.** Superimposed ultra-violet (UV) absorption spectra of standard caffeine and paracetamol ‐ and caffeine and paracetamol in PANEXT and SAFEXT.

At a wavelength of 260 nm, the maximum densitometric responses of caffeine and paracetamol in standards and the commercial tablets PANEXT and SAFEXT were recorded. The specificity of the greener HPTLC technique for the SMD of caffeine and paracetamol was demonstrated by the identical UV spectra, R<sup>f</sup> data, and wavelengths of caffeine and paracetamol in standards and the commercial tablets PANEXT and SAFEXT.

#### *2.3. Application of Greener Normal-Phase and Reversed-Phase HPTLC Aapraches in the SMD of Caffeine and Paracetamol in Commercial Tablets*

For the SMD of caffeine and paracetamol in their commercial formulation, the greener normal-phase and reversed-phase HPTLC techniques were used as an alternative to regular analytical approaches. The chromatograms of caffeine and paracetamol from commercial tablets were identified by comparing the TLC spots at R<sup>f</sup> = 0.40 ± 0.01 for caffeine and R<sup>f</sup> = 0.59 ± 0.02 for paracetamol in comparison with those of standards for caffeine and paracetamol using the greener normal-phase HPTLC approach. Figure 6 summarizes the recorded chromatograms of caffeine and paracetamol in the commercial tablets PANEXT (Figure 6A) and SAFEXT (Figure 6B), which showed identical peaks of caffeine and paracetamol to those of standards for caffeine and paracetamol in both the commercial tablets.

The chromatograms of caffeine and paracetamol from commercial tablets were identified by comparing their TLC spots at R<sup>f</sup> = 0.43 ± 0.01 for caffeine and R<sup>f</sup> = 0.57 ± 0.02 for paracetamol with those of standards caffeine and paracetamol using the greener reversedphase HPTLC approach. Figure 7 summarizes the recorded chromatograms of caffeine and paracetamol in the commercial tablets PANEXT (Figure 7A) and SAFEXT (Figure 7B), which also showed identical peaks of caffeine and paracetamol to those of standards caffeine and paracetamol in both commercial tablets.

‐ ‐

‐ ‐

‐ ‐ **Figure 6.** Normal-phase HPTLC chromatograms of caffeine and paracetamol in (**A**) commercial tablets PANEXT and (**B**) commercial tablets SAFEXT. ‐ 

*‐ ‐* 

*‐ ‐* 

> ‐ ‐ ‐

‐ ‐

‐ ‐ ‐

‐ ‐

‐ ‐ ‐

‐ ‐

‐ ‐ ‐

‐

‐

‐

‐

‐

**Figure 7.** Reversed-phase HPTLC chromatograms of caffeine and paracetamol in ( ‐ **A**) commercial tablets PANEXT and (**B**) commercial tablets SAFEXT.

‐ Using the greener normal-phase HPTLC technique, the percent assays of caffeine in the commercial PANEXT and SAFEXT tablets were estimated to be 91.23 ± 1.14 and 92.45 ± 1.22 percent, respectively. Using the greener normal-phase HPTLC technique, the percent assays of paracetamol in commercial PANEXT and SAFEXT tablets were estimated to be 89.41 ± 1.04 and 91.13 ± 1.06 percent, respectively. Using the greener reversed-phase HPTLC technique, the percent assays of caffeine in commercial PANEXT and SAFEXT tablets were estimated to be 98.51 ± 1.42 and 101.12 ± 1.53 percent, respectively. Using the greener reversed-phase HPTLC technique, the percent assays of paracetamol in commercial PANEXT and SAFEXT tablets were estimated to be 99.42 ± 1.45 and 100.64 ± 1.49 percent, respectively. The greener normal-phase and reversed-phase HPTLC methods were shown to be suitable for the SMD of caffeine and paracetamol in commercial formulations. However, for the SMD of caffeine and paracetamol in commercial formulations, the reversedphase HPTLC methodology was more reliable than the normal-phase HPTLC methodology.

#### *2.4. Greenness Estimation Using AGREE*

Various quantitative approaches are available for the greenness estimation of analytical approaches [36–40]. However, only AGREE applies all twelve GAC principles for greenness estimation [38]. As a result, the greenness of the greener normal-phase and reversedphase HPTLC approaches was estimated by "AGREE: The Analytical Greenness Calculator (version 0.5, Gdansk University of Technology, Gdansk, Poland, 2020)". The typical diagram for the AGREE scale of the greener normal-phase and reversed-phase HPTLC techniques is shown in Figure 8. The AGREE scale was estimated to be 0.81 and 0.83 for the greener normal-phase and reversed-phase HPTLC methods, respectively. These findings indicated the excellent greenness nature of the greener normal-phase and reversed-phase HPTLC approaches for the SMD of caffeine and paracetamol in their commercial formulations.

‐

‐ ‐

‐ ‐

‐ ‐

‐ ‐

‐ ‐

‐ ‐

‐

‐

‐

‐

‐ ‐

‐

‐ ‐

‐ ‐ **Figure 8.** "Analytical GREEnness (AGREE)" scale for (**A**) greener normal-phase HPTLC and (**B**) greener reversed-phase HPTLC methods.

#### **3. Materials and Methods**

*3.1. Materials*

‐

‐ ‐ ‐ The standards of caffeine and paracetamol were provided by "Sigma Aldrich (St. Louis, MO, USA)". HPLC-grades E and EA were provided by "E-Merck (Darmstadt, Germany)". The W was obtained from the Milli-Q unit in the laboratory. The commercial tablets PANEXT and SAFEXT were obtained from the pharmacy shop in "Al-Kharj, Saudi Arabia". All other solvents utilized were of analytical grades.

#### *3.2. Instrumentation and Analytical Procedures*

The "HPTLC CAMAG TLC system (CAMAG, Muttenz, Switzerland)" was used for the SMD of caffeine and paracetamol in their standards and commercial tablets. The sample solutions were spotted as 6-mm bands utilizing a "CAMAG Automatic TLC Sampler 4 (ATS4) Sample Applicator (CAMAG, Geneva, Switzerland)". The "CAMAG microliter Syringe (Hamilton, Bonaduz, Switzerland)" was linked with sample applicator. The application rate for the SMD of caffeine and paracetamol was fixed at 150 nL/s. Under linear ascending mode, the TLC plates were developed in a "CAMAG automated developing chamber 2 (ADC2) (CAMAG, Muttenz, Switzerland)" at a distance of 80 mm. For 30 min at 22 ◦C, the development chamber was saturated with vapors of greener mobile phases. Caffeine and paracetamol were detected using a wavelength of 260 nm. The slit size (band length × width) and scanning rate were both set at 4 mm × 0.45 mm and 20 mm/s, respectively. Three or six replicates were used for each estimation. The software used was "WinCAT's (version 1.4.3.6336, CAMAG, Muttenz, Switzerland)".

The greener normal-phase and reversed-phase HPTLC methodologies used the same instrumentation and analytical procedures as the normal-phase and reversed-phase HPTLC approaches. The TLC plates and the greener mobile phase components were found to be the most significant differences between the two procedures. In the greener normalphase HPTLC technique, the TLC plates were glass plates (plate size: 10 cm × 20 cm) pre-coated with normal-phase silica gel (particle size: 5 µm) 60F254S plates, but in the greener reversed-phase HPTLC approach, the TLC plates were glass plates (plate size: 10 cm × 20 cm) pre-coated with reversed-phase silica gel (particle size: 5 µm) 60F254S plates. In both cases, the polymer-binder plate was not used. In the greener normal-phase HPTLC approach, the greener mobile phase was EA/E (85:15, *v*/*v*); however, in the greener reversed-phase HPTLC approach, the greener mobile phase was E/W (50:50, *v*/*v*).

#### *3.3. Calibration Curves and QC Sample for Caffeine and Paracetamol*

Caffeine and paracetamol stock solutions were made individually by dispensing the requisite amounts of both molecules in the specified amount of respective mobile phase, resulting in a final stock solution of 100 µg/mL for both compounds. The concentrations in the 50–500 ng/band range for caffeine and paracetamol were generated using the greener normal-phase HPTLC methodology and the 25–800 ng/band range for caffeine and paracetamol using the greener reversed-phase HPTLC methodology by diluting variable volumes of caffeine or paracetamol stock solution with the respective mobile phase. For the normal-phase HPTLC methodology, 200 µL of each concentration of caffeine and paracetamol were put to normal-phase TLC plates and reversed-phase TLC plates for the reversed-phase HPTLC methodology. Using both analytical techniques, the spot area of each concentration of caffeine and paracetamol was measured. Caffeine and paracetamol calibration curves were created by graphing the concentrations of both drugs against the observed spot area in six repeats (*n* = 6). For the determination of various validation parameters, three distinct QC samples were prepared freshly.

#### *3.4. Processing of Samples for the SMD of Caffeine and Paracetamol in Commercial Tablets*

Ten commercial tablets (each containing 65 mg of caffeine and 500 mg of paracetamol) were weighed and the average weights were computed for the SMD of caffeine and paracetamol in PANEXT and SAFEXT. Each brand's tablets were coarsely crushed and powdered. A portion of each brand's powder was dissolved in 100 mL of the relevant mobile phase. For the greener normal-phase and reversed-phase HPTLC methods, 1 mL of this solution of each brand of tablet was diluted again using 10 mL of the corresponding mobile phase. The prepared solutions of PANEXT and SAFEXT commercial tablets were filtered and sonicated for around ten minutes to remove any undissolved excipients. Using the greener normal-phase and reversed-phase HPTLC methods, the generated solutions were used to determine caffeine and paracetamol in commercial tablets PANEXT and SAFEXT.

#### *3.5. Analytical Method Validation*

Utilizing the ICH-Q2-R1 recommendations, the normal-phase and reversed-phase HPTLC techniques for the SMD of caffeine and paracetamol were validated for various parameters [41]. By graphing the concentrations of caffeine and paracetamol against their measured spot area, the linearity range for caffeine and paracetamol was discovered. The normal-phase HPTLC approach's linearity for caffeine and paracetamol was evaluated in the 50–500 ng/band range (*n* = 6). For the reversed-phase HPTLC method, the linearity for caffeine and paracetamol was evaluated in the 25–800 ng/band range (*n* = 6).

The calculation of R<sup>f</sup> , As, and N/m was used to evaluate the parameters for the system acceptability for the greener normal-phase and reversed-phase HPTLC techniques for the SMD of caffeine and paracetamol. For both analytical approaches, the R<sup>f</sup> , As, and N/m data were computed utilizing their reported equations [39].

The percent recovery was utilized to examine the accuracy of the normal-phase and reversed-phase HPTLC methods for the SMD of caffeine and paracetamol. For caffeine and paracetamol, the accuracy of the greener normal-phase HPTLC technique was tested at three QC levels: lower QC (LQC; 100 ng/band), middle QC (MQC; 300 ng/band), and high QC (HQC; 500 ng/band). For caffeine and paracetamol, the accuracy of the greener reversed-phase HPTLC technique was tested at three QC levels: LQC (50 ng/band), MQC (300 ng/band), and HQC (800 ng/band). Using both analytical techniques, the percent of recovery for caffeine and paracetamol (*n* = 6) was assessed at each QC level.

Intra/inter-assay precision was measured for the greener normal-phase and reversedphase HPTLC methods for caffeine and paracetamol. Quantitation of newly prepared caffeine and paracetamol solutions at LQC, MQC, and HQC on the same day for both analytical techniques (*n* = 6), was used to examine intra-assay variation for caffeine and paracetamol. Quantitation of freshly prepared solutions at LQC, MQC, and HQC on three

consecutive days for both analytical techniques (*n* = 6) was used to investigate inter-assay variation for caffeine and paracetamol.

For both analytical techniques, the robustness for caffeine and paracetamol was evaluated by making some slight purposeful modification in the mobile phase composition. The greener mobile phase EA/E (85:15, *v*/*v*) for caffeine and paracetamol was altered to EA/E (87:13, *v*/*v*) and EA/E (83:17, *v*/*v*) for the greener normal-phase HPTLC technique, and the variations in chromatographic response and R<sup>f</sup> values were recorded (*n* = 6). The greener mobile phase E/W (50:50, *v*/*v*) for caffeine and paracetamol was altered to E/W (52:48, *v*/*v*) and E/W (48:52, *v*/*v*) for the greener reversed-phase HPTLC technique, and the variations in chromatographic response and R<sup>f</sup> values were recorded (*n* = 6).

By using a "standard deviation" technique, the sensitivity of the greener normal-phase and reversed-phase HPTLC approaches for caffeine and paracetamol was examined as "LOD and LOQ". Caffeine and paracetamol "LOD and LOQ" were computed using their published equations for both analytical procedures (*n* = 6) [41].

The R<sup>f</sup> values and UV spectra of caffeine and paracetamol in commercial tablets PANEXT and SAFEXT were compared with those of standards caffeine and paracetamol to determine the specificity of the greener normal-phase and reversed-phase HPTLC methods for caffeine and paracetamol.

#### *3.6. Application of Greener Normal-Phase and Reversed-Phase HPTLC Approaches in the SMD of Caffeine and Paracetamol in Commercial Tablets*

For the normal-phase HPTLC technique, the obtained solutions of the commercial tablets PANEXT and SAFEXT were put on normal-phase TLC plates and on reversed-phase TLC plates for the reversed-phase HPTLC technique. For all analytical techniques, the chromatographic responses were documented using the identical experimental circumstances employed for the SMD of standards caffeine and paracetamol (*n* = 3). For both analytical procedures, the quantities of caffeine and paracetamol in commercial tablets were approximated using the calibration curves for caffeine and paracetamol.

#### *3.7. Greenness Estimation Using AGREE*

The AGREE technique [38] was utilized to assess the greenness scale for the normalphase and reversed-phase HPTLC procedures for the SMD of caffeine and paracetamol. The AGREE scales (0.0–1.0) for the greener normal-phase and reversed-phase HPTLC approaches was estimated utilizing "AGREE: The Analytical Greenness Calculator (version 0.5, Gdansk University of Technology, Gdansk, Poland, 2020)" for both the analytical approaches.

#### **4. Conclusions**

The literature lacks greener analytical techniques for the SMD of caffeine and paracetamol. As a result, compared to the greener normal-phase HPTLC approach, this research was carried out to develop and validate the rapid, sensitive, and greener reversed-phase HPTLC approach for the SMD of caffeine and paracetamol in their commercial tablets. For the SMD of caffeine and paracetamol, the greener reversed-phase HPTLC approach is more linear, accurate, precise, robust, and sensitive than the greener normal-phase HPTLC approach. The quantities of caffeine and paracetamol in commercial tablets PANEXT and SAFEXT were found to be significantly higher using the reversed-phase HPTLC methodology compared with the normal-phase HPTLC methodology. The AGREE estimation showed the excellent green properties of both the analytical approaches. For the SMD of caffeine and paracetamol in commercial formulations, the greener reversed-phase HPTLC approach has been presented superior to the greener normal-phase HPTLC approach based on different validation criteria and pharmaceutical assays.

**Author Contributions:** Conceptualization, supervision—P.A. and F.S.; Methodology—A.I.F., P.A., M.H.A. and T.M.A.; Validation—A.A. (Abuzer Ali), S.A. and F.S.; Data curation—A.A. (Amena Ali), F.K.A. and M.M.G.; Funding acquisition—A.A. (Abuzer Ali); Project administration—P.A.; Software— P.A., S.A. and F.S.; Writing original draft—F.S.; Writing—review and editing—A.A. (Abuzer Ali), M.M.G. and S.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Taif University Researchers Supporting Project (Number TURSP-2020/124), Taif University, Taif, Saudi Arabia. The APC was funded by TURSP.

**Acknowledgments:** Authors are thankful to the Taif University Researchers Supporting Project (Number TURSP-2020/124), Taif University, Taif, Saudi Arabia for supporting this work.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Sample Availability:** Samples of the caffeine and paracetamol compounds are available from the authors.

#### **References and Note**


## *Article* **Application of Skyline for Analysis of Protein–Protein Interactions In Vivo**

**Arman Kulyyassov**

Republican State Enterprise "National Center for Biotechnology" under the Science Committee of Ministry of Education and Science of the Republic of Kazakhstan, 13/5, Kurgalzhynskoye Road, Nur-Sultan 010000, Kazakhstan; kulyyasov@biocenter.kz; Tel.: +7-7172-707534

**Abstract:** Quantitative and qualitative analyses of cell protein composition using liquid chromatography/tandem mass spectrometry are now standard techniques in biological and clinical research. However, the quantitative analysis of protein–protein interactions (PPIs) in cells is also important since these interactions are the bases of many processes, such as the cell cycle and signaling pathways. This paper describes the application of Skyline software for the identification and quantification of the biotinylated form of the biotin acceptor peptide (BAP) tag, which is a marker of in vivo PPIs. The tag was used in the Proximity Utilizing Biotinylation (PUB) method, which is based on the co-expression of BAP-X and BirA-Y in mammalian cells, where X or Y are interacting proteins of interest. A high level of biotinylation was detected in the model experiments where X and Y were pluripotency transcription factors Sox2 and Oct4, or heterochromatin protein HP1γ. MRM data processed by Skyline were normalized and recalculated. Ratios of biotinylation levels in experiment versus controls were 86 ± 6 (3 h biotinylation time) and 71 ± 5 (9 h biotinylation time) for BAP-Sox2 + BirA-Oct4 and 32 ± 3 (4 h biotinylation time) for BAP-HP1γ + BirA-HP1γ experiments. Skyline can also be applied for the analysis and identification of PPIs from shotgun proteomics data downloaded from publicly available datasets and repositories.

**Keywords:** biotin acceptor peptide (BAP); biotin ligase BirA; liquid chromatography tandem mass spectrometry (LC-MS/MS); multiple reaction monitoring (MRM); protein–protein interactions (PPIs); proximity utilizing biotinylation (PUB); proteomics

#### **1. Introduction**

Wide practical application of liquid chromatography in combination with mass spectrometry has been observed recently in proteomics [1,2] and metabolomics [3,4] as a routine method for the qualitative and quantitative analysis of biological samples. For example, when optimizing expression, performing quality control, or studying pharmacokinetics of recombinant proteins, it is crucial that the best conditions for production or analysis of the drug products are found [5,6]. Another important task is to obtain information about changes in the expression of marker proteins under different physiological conditions of the cell [7]. Examples of this include: differences in protein composition in a healthy/cancer cell or differences under the influence of external factors such as temperature, chemical agents, or radiation provide valuable information about metabolic and signaling pathways, mechanisms of stress response. In all these cases, the results are obtained as chromatograms in the multiple reaction monitoring (MRM) method, where many peptides derived from target proteins can be identified by retention time and mass spectra of fragment ions (or MS/MS spectra), and the relative amount of each peptide between samples can be determined by comparison of the peak areas [8–10].

However, information on protein composition is not sufficient to fully understand the mechanism of cell function. The quantification of protein–protein interactions (PPIs) in vivo can be a useful extension in research since more than 80% of proteins do not function separately, but rather interact and participate in the formation of stable or transient

**Citation:** Kulyyassov, A. Application of Skyline for Analysis of Protein–Protein Interactions In Vivo. *Molecules* **2021**, *26*, 7170. https:// doi.org/10.3390/molecules26237170

Academic Editors: Victoria Samanidou and Natasa Kalogiouri

Received: 2 November 2021 Accepted: 23 November 2021 Published: 26 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

complexes [11]. These protein–protein interactions play an important role in almost all vital processes in cells, such as DNA replication, gene transcription and translation, signal transduction, cell-cycle control and proliferation, and cell–cell communication [12].

Methods based on a combination of affinity purification (AP) or tandem affinity purification (TAP) and mass spectrometry (MS/MS) have now become standard for the identification of protein partners [13–16]. However, these methods have the serious drawback of a large number of false-positive identifications [17]. In addition, the cell lysis procedure can lead to the destruction of weak protein–protein interactions, which can also lead to false-negative results. For example, the list of Oct4-interacting proteins identified by co-immunoprecipitation (Co-IP) did not include one of the most studied Oct4 partners, namely Sox2 [18].

Enzymatically catalyzed proximity labeling is an alternative to immunoprecipitation and biochemical fractionation for the proteomic analysis of macromolecular complexes and protein interaction networks [19]. In this method, ligation enzymes are expressed in cells as conjugates with proteins of interest. For example, proximity-dependent biotinylation methods are based on the use of mutant biotin ligases, BioID [20] or TurboID [21]. These BirA mutants prematurely release the highly reactive yet labile biotinoyl-AMP inside of a living cell, which readily reacts with lysine's primary amino groups of proximal proteins.

On the other hand, the proximity utilizing biotinylation (PUB) method is based on the use of humanized wild-type biotin ligase BirA fusions. The wild-type BirA uses biotin and ATP to generate biotinoyl-AMP [22,23]. Wild-type BirA holds on to the reactive biotin molecule until it is covalently attached to a very specific substrate called biotin acceptor peptide (BAP) [24]. Thus, biotinylation is a result of the direct contact of BirA and BAP parts of recombinant proteins which occurs in cases of protein–protein interaction or random collision in vivo.

Both methods are based on similar principles; this is the in vivo creation of a permanent covalent mark on one of the proteins of interest or partners interacting with them, which allows us to bypass the limitations imposed by the extraction and purification stages. Ultimately, results will be obtained with much fewer false-positive and false-negative protein identifications compared to traditional methods such as IP-MS/MS or TAP. The result is the facilitation of the bioinformatic part of data analysis.

The aim of this work was to use the Skyline program to process the results of experiments on the quantitative analysis of PPI using the proximity utilizing biotinylation (PUB) method.

#### **2. Results and Discussion**

#### *2.1. Overview of the Method and Experimental Workflow of the PUB Protocol*

The principle of the PUB method is based on using enzyme/substrate pair reactions [25–28], where two proteins to be tested for their interaction in vivo are co-expressed in mammalian cells, one as fused to the BAP, and the other fused to an enzyme BirA, which is an *Escherichia coli* protein biotin ligase [29]. When the two proteins are in proximity to each other, for example, when an interaction of X and Y occurs in vivo, a more efficient biotinylation of the BAP is to be expected (Figure 1A). The biotinylation status of the BAP fusion protein can be further monitored by Western blot, mass spectrometry, or confocal microscopy (Figure 1B,C). HEK293T, HeLa, or MRC-5 fetal lung fibroblast cell lines can be used for the transient or stable expression of recombinant proteins in the PUB method. Usually, one control experiment is performed in a parallel dish or a 6-well plate, using cells in which non-interacting proteins or other pair proteins are expressed for comparison (BirA-X and BAP-Z). Depending on the proteins of interest chosen for the experiment, biotin is added to the medium from 5 min to 9 h before harvesting the cells. The sequence of the BAP peptide was modified and compared to commonly used peptides, such as Avitag [30,31], in order to reduce the level of background biotinylation [28]. Additionally, 7His-tag was added upstream of the sequence to provide the option to purify both labeled and nonlabeled BAP fusion proteins from cell lysates.

**Figure 1.** Principle of PUB method in living cell and workflow for quantification of PPI. In vivo interaction (**A**) of proteins X and Y results in site-specific biotinylation of the biotin acceptor peptide (BAP) by wild-type humanized biotin ligase (BirA). Biotinylated protein can be detected by WB, for example, Streptavidin-HRP, IF confocal microscopy (**B**) or LC-MS/MS (**C**). X or Y–HP1(α,β,γ), Tap54(α,β), Sox2, Oct4, or other proteins. Biotinylation levels of BAP peptides obtained after processing the results of sample analyses using Skyline (**C**) are recalculated in Microsoft Excel (**D**). P—propionylated form of BAP1070: GHHHHHHHGLTR**ILEAQK(Prop)IVR**GG, B—biotinylated form of BAP1070: GHHHHHHHGLTR**ILEAQK(Biot)IVR**GG, the sequence corresponding to the peptide on the chromatogram after trypsin digest is marked in bold. The relative ionization coefficient of tryptic peptides derived from propionylated and biotinylated BAP (**E**).

> The experimental workflow for the LC-MS/MS analysis of samples includes additional steps, such as the purification of Ni agarose beads, propionylation, and on-gel (or on-bead) tryptic digest (Figure 1C). Propionylation was used to protect the nonbiotinylated BAP peptide from tryptic cleavage on the target lysine. This modification resulted in the production of modified and nonmodified peptides of comparable sizes, facilitating the interpretation of results. After analysis on Skyline (Figure 1D), data were exported as CSV files and processed using Microsoft Excel for the calculation of biotinylation levels (Supplementary datas S1 and S2). First, the total amount of BAP was calculated by the addition of the total area of propionylated BAP to the total area of recalculated biotinylated BAP. For the recalculation of the biotinylated BAP, the relative ionization coefficient k = 11.9 was used (Figure 1E), which was estimated earlier in SILAC experiments [28]. The ionization efficiency depends on the chemical structure of a molecule and would thus be different for propionyl and biotin residues. Therefore, a direct comparison between total ion chromatograms (TIC) of the biotinylated and propionylated BAP in LC-MS/MS data is not possible. After this step, the areas were normalized, and normalization coefficients of the total amounts of BAP were calculated for each sample. These normalization coefficients

α β γ α β

were then used to recalculate the biotinylation levels and for the estimation of means and standard deviations.

The total amount of BAP was calculated using the formula ABAP = k × AbBAP + ApBAP, where AbBAP corresponds to the peak area of total ion chromatograms (TIC) of biotinylated, and ApBAP to propionylated BAP, and k is the relative ionization coefficient between the biotinylated and propionylated BAP peptides (Figure 1D). The chromatographic elution peaks of the fragments for the four most intensive ions, *y*7, *y*6, *y*5, and *y*4, in extracted ion chromatograms (EIC) were integrated and summed to give the peak area of TIC. In the control (BAP-Y + BirA-Z) and in the experiment (BAP-Y + BirA-X), as well as in the replicates, the expression levels of recombinant BAP-Y proteins may differ. Variations in the total amount of BAP-Y can also appear during a sample preparation. Thus, direct comparison of the biotinylation levels AbBAP between samples is not correct. Therefore, the total amount of BAP-Y, including its biotinylated and propionylated forms in all samples, was normalized and the normalization coefficients were determined (Supplementary data S2, Table S1).

The Skyline is an application for targeted proteomics and quantitative data analysis in the frame of the Windows operation system [32]. Its interface facilitates the improvement of mass spectrometer methods and the analysis of data from targeted MRM experiments. Skyline imports the native output files from instruments manufactured from different vendors smoothly, connecting mass spectrometer output back to the experimental design document. A rich choice of graphics displays provide powerful tools for inspecting and monitoring data integrity as data are acquired, helping instrument operators to identify problems early. It is open-source and freely available for commercial and academic use [9,33]. In addition, its output data format (csv.files) allows the performance of post-processing analysis in Microsoft Excel to recalculate biotinylation levels.

This software was successfully used to identify and quantify the target BAP peptides from all MRM data (Supplementary data S1, Figures S1 and S2). Since the amino acid sequence of the BAP1070 peptide is artificially generated and is absent in the NCBI and Swissprot databases, the sequence of this peptide was added to a client-made database, BAP1070, using the Database manager (Figure 2A). This allowed the DAT file to be generated and the spectral libraries to be created in Skyline.

**Figure 2.** Creation of BAP1070 file. (**A**) Screenshots of the database manager page with BAP1070 file, (**B**) Mascot search results and exporting the DAT file.

Prior to importing raw data into Skyline, a spectral library containing the product ion spectra of the BAP target peptides was constructed using the DAT file. The spectral library consisted of MS/MS spectra of biotinylated and propionylated forms of BAP1070 peptide. A spectral library allowed for the direct comparison of BAP target peptide product ion spectra from the MRM analyses to the corresponding product ion "library match". Product ion transitions used to confirm the identity of each target peptide in the MRM analyses were automatically picked based on the four most abundant y-type product ion intensities observed in the "library match" spectrum.

#### 2.1.1. Creation of MRM Method and LC-MS/MS Analysis of the Samples

The vendors' default method (Bruker Company) was used for the creation of the MRM method, as described earlier [34]. Precursor ions: *m*/*z* 563.2 (ILEAQK(Prop)IVR) propionylated form of BAP, and *m*/*z* 648.8 (ILEAQK(Biot)IVR) biotinylated form of BAP.

#### 2.1.2. Creation of BAP1070 Database on Mascot Search Server Using Database Manager

Before analysis on Skyline, the raw LC-MS/MS data were processed to a special format—DAT file. First, the raw data were analyzed on Bruker DataAnalysis software to generate an MGF file. Then, the BAP1070 database was created containing a sequence of this peptide in the Database manager (Figure 2A), which is a browser-based utility for updating and configuring local copies of sequence databases. Analysis of an MGF file against the BAP1070 database in the Mascot search engine, including propionylation and biotinylation modifications, yielded a report where results could also be exported as a DAT file (Figure 2B).

#### 2.1.3. MRM Analysis and Post-Processing of Data

All MRM data were analyzed in Skyline 19.1.0.193. A spectral library was constructed from the peptide identifications from a DAT file exported from the Mascot result page. The four product ions extracted by Skyline were determined based on the ranking of the top four most intense y-ions from the corresponding library spectrum for each peptide. Dot-product (dotp) scores were calculated based on the correlation of the measured product ion peak intensities with the peak intensities observed in the library spectrum for that same peptide [33]. Raw LC-MS/MS data and processed files were uploaded to the Panorama repository [35].

#### *2.2. DNA Dependent Interaction of Sox2 and Oct4*

After processing and recalculating Skyline results, quantitative data on biotinylation levels were obtained (Figure 3A,B). The samples from experiments with the co-expression of BAP-Sox2 and BirA-Oct4 in HEK293T cells showed a high level of biotinylation (86 ± 6 and 71 ± 5 for different biotinylation times). This is due to the presence of DNA binding domains, HMG present in Sox2 and POU in Oct4, which recognize *Utf1* or other motifs [36] and result in close contact between target BAP and biotin ligase BirA. Contrary to Sox2, GFP lacks DNA binding domains, and in a control experiment with the coexpression of BAP-GFP and BirA-Oct4, very low biotinylation levels were observed (sample 0–9 on Figure 3A). Recombinant proteins BAP-GFP and BAP-Sox2 from HEK293T cell nuclear lysates were purified on Ni sepharose beads and propionylated before trypsin digest, as described earlier [34].

**Figure 3.** Skyline graphical representation of chromatograms from PUB experiments. Total ion chromatograms of propionylated (P in green circle) and biotinylated (B in red circle) BAP peptides for two examples of experimental PPIs (**B**,**D**) obtained using different instrument platforms (**A**–**B**, **C**–**D**). Four most intense fragment ions, *y*<sup>7</sup> , *y*<sup>6</sup> , *y*<sup>5</sup> and *y*<sup>4</sup> , were chosen for area calculation of biotinylated BAP peptide in extracted ion chromatograms. Left side are controls BAP-GFP + BirA-Oct4 (**A**) and BAP-HP1γ + BirA-Tap54α (**C**) and right side of the figure represents experiments with interacting proteins—BAP-Sox2 + BirA-Oct4 (**B**) and BAP-HP1γ + BirA-HP1γ (**D**). The average ratios of biotinylation levels were obtained from three experiments after recalculation and normalization.

#### *2.3. Protein Oligomerization HP1γ-HP1γ*

Heterochromatin protein HP1γ was chosen as another example of protein–protein interactions. The proteins of this family contain a chromo shadow domain (CSD), which allows them to form dimers and oligomers [37,38]. The formation of these oligomeric structures is critical for the organization of heterochromatin in the cell nucleus [39]. In the experiment, BAP-HP1γ and BirA-HP1γ protein pairs were expressed. Another protein, TAP54α, participates in the formation of hexamers with ATPase activity and is a component of histone acetyltransferase complexes [40,41]. Since Tap54α is not a protein that interacts with HP1γ, we chose a model where other protein pairs, BAP-HP1γ and BirA-Tap54α, were expressed in a separate dish as a control. The difference in the biotinylation level of the experiment (BAP-HP1γ + BirA-HP1γ) versus control (BAP-HP1γ + BirA-Tap54α) after processing the raw mass spectrometer data with Skyline was also significant and was 32 ± 3 (Figure 3C,D). The raw data were obtained on an Agilent nanoHPLC-Chip-3D6340 Iontrap instrument and converted to mzML format [42] using ProteoWizard software [43,44].

#### *2.4. Analysis of Shotgun Proteomics Data (Mutant Biotin Ligase BioID Application)*

The Skyline program was mainly developed for the analysis of MRM results in targeted proteomics [32]. However, this application can also be used to analyze the results of Shotgun proteomics where an instrument is operated in data-dependent acquisition (DDA) mode. For validation, the raw data from the results of the BioID mutant biotin ligase experiments, published recently by Go et al. [45] and publicly available in the Massive repository, were downloaded. Go et al. identified 35,902 interactions with 4424 unique high-confidence proximity interactors for 192 BioID fused bait proteins from different cellular compartments. The MGF file was used to obtain the DAT file, as described earlier in Section 2.1.2, which was used in Skyline to build the spectral library. Since these results were obtained on a different instrument platform (Eksigent NanoLC-Ultra 2D plus HPLC system-Orbitrap Elite) and under different modes of operation, the parameters in the Peptide setting and Transition settings tabs were changed, as described in the experimental part from paper [45]. An example of interacting (or proximal) proteins of mitochondrial Pyruvate Dehydrogenase E1 Subunit Alpha 1 (PDHA1) fused with BioID is shown in Figure 4.

#### *2.5. Perspectives of Enzymatically Catalyzed Labeling for Biological Research*

Methods based on the use of wild-type biotin ligase BirA and their mutant versions, BioID or TurboID, have a similar principle: the creation of a permanent covalent label on the partner protein in vivo. This facilitates subsequent steps in protocols and especially easy and efficient purification of biotinylated proteins from cell lysates using commercially available reagents and kits.

In addition, these methods can be complementary to each other. For example, while the use of BioID or TurboID allows the identification of proximal or partner proteins in cells, the PUB method can be used to quantitatively compare the identified protein–protein proximity.

The MRM method, where the output results are presented as coupled data of chromatographic parameters and mass spectra for each peptide, is now widely used to study the mechanisms of external influences (temperature, radiation, or chemical reagents) on the expression of various proteins in a cell. By analogy, the PUB method can be used to study the influence of various external factors on protein–protein interactions in a living cell, which can extend a given research area and provide additional information about cell organization and function.

Thus, the use of modern bioinformatics programs such as Skyline in combination with PUB, BioID, and TurboID methods will facilitate the analyses of large amounts of data to solve various problems of cell and molecular biology.

**Figure 4.** Results of analysis of data from paper Go et al. [45] using Skyline program. All peptides are grouped into lists from interacting or proximal proteins on the left side, the total ion chromatogram (TIC) in the center and MS/MS spectra are shown in the right part of this figure. Example of ATP synthase (inset on left side) as a protein, proximal to PDHA1.

#### **3. Materials and Methods**

Cell culture, transient transfection, and sample preparation steps were described earlier [28,34].

The peptide mixtures were analyzed using two LC-MS/MS systems:


The LC-MS/MS instruments were set to monitor transitions of biotinylated (*m*/*z* 648.8, collision energy 33.0 eV) and propionylated (*m*/*z* 563.2, collision energy 27.0 eV) forms of BAP peptide in samples.

#### *3.1. Data Preparation and Creation of MGF File Using DataAnalysis*

For preliminary analysis of data and generation of peak lists, DataAnalysis (DA 4.1) software was used. Retention-time information was changed to seconds.

The MGF file was generated from raw data by clicking the following tabs on the menu: Find/Compounds MS(n)→Deconvolute/Mass spectra → File/Export/Compounds. Subsequent database searches were performed using Mascot search engine. Then, the results were imported as the DAT file which were used to build a spectral library in Skyline.

*3.2. Creation of BAP1070 Fasta Database on Mascot Search Server*

In MS Notepad text editor, the aminoacid sequence of BAP1070 was pasted with the description line as follows:

>BAP1070

GHHHHHHHGLTRILEAQKIVRGG

This file was saved as BAP1070\_fasta.txt

On mascot server http://mascot-server/mascot/index.html (Configuration last updated Thu Apr 15 10:36:27 2021), the BAP1070 database was created using the following steps: Home subpage → MascotUtilities → ConfigurationEditor → Database Manager → Fasta → Create new. Configuration details for BAP1070: Database name—BAP1070, Database type—aminoacid, Accession parse rule— > [ˆ] \*\(.\*\), Description parse rule > [ˆ] \*\(.\*\), Taxonomy source—none, Sequence report source—FASTA file, Full-text report source—None, Number of threads—automatic, Use memory mapping?—Yes, Lock to memory?—No.

Analysis of data, including Building a Spectral Library in Skyline, Configuring Transition Settings, Populating the Skyline Peptide Tree, Importing Raw Data into Skyline and Subsequent Filtering, and data processing and calculation of biotinylation levels are described in Supplementary data S1, Figures S1 and S2.

#### **4. Conclusions**

In this study, the Skyline program was used for the first time to analyze results obtained by using a proximity utilizing biotinylation method based on expression in mammalian target cells BAP-X and wild-type BirA-Y protein conjugates (first example: X-Sox2, Y-Oct4, versus control X-GFP, Y-Oct4 and second: example X, Y-HP1γ versus control X-HP1γ, Y- Tap54α). Peak areas of biotinylated BAP were used for the estimation of PPI, while peak areas of propionylated BAP on MRM chromatograms were used for the recalculation and normalization of data between different samples. This program allowed for fast processing of raw data, the calculation of peak areas, and provided the output file in CSV format, which is convenient for subsequent analysis on Microsoft Excel.

Skyline was also used to analyze data on protein–protein interactions and proximities obtained by using mutant biotin ligase BioID [45]. These raw data were downloaded from the MassIVE Repository database and were sourced from another LC-MS/MS instrument platform, demonstrating that the Skyline program is not "instrument or vendor-oriented".

Overall, the Skyline program offers an advantage in that it provides a good graphic representation of data and reduces analysis time. This protocol could be applicable, not only to BAP, but also to other synthetic peptides which are absent in NCBI or SwissProt databases.

**Supplementary Materials:** The following are available online, Figure S1: Extracted ion chromatograms (EICs) of the top four ranked y-ions, Figure S2: Relative quantification diagrams for biotinylated and propionylated forms of BAP after processing on Skyline Table S1: Calculation of corrected biotinylation levels.

**Funding:** This research was funded by a grant from the Ministry of Education and Science of the Republic of Kazakhstan AP09259838 "Application of new proteomics methods in studying the mechanism of action of pluripotency transcription factors expressed in mammalian cell lines" for 2021-2023 (State registration number 0121PK00163).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The MS proteomics data have been deposited into the ProteomeXchange Consortium via the PRIDE partner repository, with the data set identifier PXD015756.

**Acknowledgments:** A.K. is greatly thankful to Yerlan Ramankulov (National Center for Biotechnology, Kazakhstan) for funding acquisition, Ruslan Kalendar (Department of Agricultural Sciences, University of Helsinki, FI-00014 Helsinki, Finland,) for help in reading and editing this manuscript, and Madina Zhunusova (National Center for Biotechnology, Kazakhstan) for assistance with HEK293T cell culture.

**Conflicts of Interest:** The author declares no conflict of interest.

**Sample Availability:** The vector plasmids, pcDNA3-BAP-Sox2 and pOz-humBirA-GFP, for transient transfection in cells are available from Addgene (Addgene ID 133281 and 133283, respectively).

#### **References**

