Development of an On-Line Two-dimensional Normal Phase Liquid Chromatography System for Analysis of Weakly Polar Samples

In this study, a novel on-line two-dimensional (2D) normal phase × normal phase liquid chromatography (NPLC × NPLC) was developed for the separation of weakly polar samples. The 2D NPLC was integrated by a unique designed solvent evaporation (SE) interface, including a 6-port 2-position valve and a 10-port 2-position valve in conjunction with two silica gel-packed enrichment columns. The enrichment columns played a pivotal role in evaporating normal phase (NP) solvent from the first dimensional NPLC under vacuum and elevated temperature condition. The working parameters of the interface were evaluated comprehensively. To demonstrate the resolving powerful 2D system, we analyzed three natural-sourced weakly polar compounds from the extracts of toad venom, dammar resin, and propolis. To this end, we studied the effects of mobile phase combinations, the second dimensional column length for separation of the extracts from toad venom, and propolis, respectively. In general, we have found that excellent separation can be obtained with diversity of NP solvent combinations and longer NPLC column (150 mm long). Moreover, compared with the generally used 2D reversed-phase liquid chromatography (RPLC × RPLC), the NPLC × NPLC method exhibited better separation orthogonality. In conclusion, the new NPLC × NPLC separation method provides potential advantages for analysis of weakly polar samples.


Introduction
Complex samples derived from natural products are always having a wide range of various physical properties and chemical polarities. Two-dimensional liquid chromatography (2D-LC) systems are capable of reducing component overlapping, which are extensively practiced in analysis of complex samples [1]. Owing to the high column efficiency and resolution, reversed-phase liquid chromatography (RPLC) is the most versatile system selected as the second dimension in various 2D-LC systems [2]. With RPLC in the second dimension, different 2D-LC combinations have been developed and applied in analysis of complicated samples [3][4][5]. Although RPLC is capable of fulfilling most separation tasks, it has weaknesses for the separation of weak polar samples. First, the excessive retention of RPLC for weakly polar samples may result in sample loss [6]. Second, the separation mechanism of hydrophobic interaction is less selective for these weakly polar compounds. In contrast, moderate retention of NPLC on weakly polar components greatly reduces the sample loss caused by irreversible adsorption. As we all know, Supercritical Fluid Chromatography (SFC) is also suitable for the separation of weakly polar compounds. The recent progress of SFC technology are in relation to chiral separations in bioanalysis [7,8]. Compared with 2D NPLC × NPLC, SFC is more expensive and complicated to operate [9]. Additionally, NPLC has more mobile phase combinations to choose, from which can provide higher separation selectivity. To the 1 3 best of our knowledge, until now, there is no report on 2D normal phase chromatography. Inevitably, however, NPLC also comes with several issues, such as the use of toxic and non-environmentally friendly solvents.
In this study, a novel on-line comprehensive 2D NPLC × NPLC system has been developed for separation of weakly polar samples. An innovative solvent evaporation (SE) interface was designed to eliminate the solvent effect caused by the eluent from the first dimensional NPLC (detailed in Supporting Information Figure S1 ~ S4). The SE interface with reference to the TEEA interface [10] was consisted by a 6-port 2-position valve and a 10-port 2-position valve, along with two silica gel-packed enrichment columns (Fig. 1). The enrichment column is the core portion of the interface, it operated under the vacuum and elevated temperature condition. First, the eluent from the first dimension flowed into the NP enrichment column, allowing rapid evaporation of the solvent and enrichment of the solute in the NP enrichment column. Second, the residue was eluted by the NP mobile phase in the second dimension for further separation. A mixture of eight reference compounds was used to investigate the parameters of the SE interface and to establish the working principle of the 2D NPLC × NPLC system. And then, the novel 2D-LC system was exploited in the separation of three natural-sourced weakly polar samples: toad venom, dammar resin, and propolis. Changing the combination of solvent system into two dimensions could achieve better separation efficiency since the diversity and flexibility of NP solvent combinations. A comparison with NPLC × NPLC and RPLC × RPLC was investigated; consequently, the novel 2D-LC method exhibited better separation orthogonality. Moreover, thanks to the low backpressure of NPLC, longer NPLC column can be used on-line in the second dimension to achieve better separation efficiency. In this report, we designed a practical and effective 2D NPLC × NPLC method development approach for the analysis of weakly polar samples.

Instrumentation
The first dimensional NPLC was performed on an Agilent 1200 analytical HPLC system, containing a quaternary pump and a DAD detector. The second NPLC system consists of a binary pump, a DAD detector, and an integrated SE interface (Fig. 1a). The SE interface includes two Agilent 1290 Infinity Thermostatted column compartments, whose highest column temperature could reach 99 °C. A 2-position 10-port switching valve (Valve A) and a 2-position 6-port switching valve (Valve B) were furnished in one of the Thermostatted column compartments for switching the flow path. Agilent Chemstation Rev. B.04.03 was used for data acquisition and processing. 2D-LC chromatograms were visualized using ACD/Spectrus 2016 (Advanced Chemistry Development Inc., Toronto, Canada).

Sample Preparation
The reference compounds were mixed in EtOH to afford the reference compounds solution with concentrations as shown in Fig. 2a.
Ground toad venom, dammar resin, and propolis (1.0 g) were extracted with EA (20 mL) for 1.5 h using Soxhlet extractor, respectively. Each solution was concentrated under vacuum at 40˚C to obtain the corresponding extract. Then, the three extracts were dissolved in CHCl 3 /MeOH (10:1) to obtain the test solution at concentration of 20 mg/mL, respectively.

SE Interface Setup
As shown in Fig. 1a, the two NPLC systems were coupled through the optimized SE interface, which was applied to eliminate the NP solvents and adsorb the solute. In the SE interface, two equivalent Cosmosil silica gel enrichment columns (10 mm × 4.6 mm i.d., 5 µm) were furnished in the 2-position 10-port switching valve (Valve A) in one of the Agilent 1290 Thermostatted column compartment. In the NP enrichment column, the elevated column temperature was up to 99 °C, and vacuum condition was applied to remove the NP solvent. The 2-port 6-position valve (Valve B) was equipped in another Thermostatted column compartment, which was designed not only to introduce the second dimensional NP mobile phase to the NP enrichment column (position A), but also to regenerate the NP enrichment column. Under optimal conditions, the recovery rate of the reference compounds ranged from 85 to 98% (in Table S2 in Supporting Information). It illustrates that moderate retention of NPLC on weakly polar components greatly reduces the sample loss caused by irreversible adsorption.

On-line 2D NPLC × NPLC Separation
After optimization of SE interface, 20 µL of toad venom, dammar resin, and propolis test solutions were injected and analyzed on the 2D NPLC × NPLC system, respectively.
An HPLCONE silica gel column (250 mm × 4.6 mm i.d., 5 µm, Microwants Co. Ltd., Suzhou, China) was used for the first dimension. In the second dimension, a Cosmosil silica gel column (50 mm × 4.6 mm i.d., 5 µm, Nakalai Tesque Co. Ltd., Kyoto Japan) or a Luna silica gel column (150 mm × 4.6 mm i.d., 3 µm, Phenomenex Co. Ltd.) was chosen. In the SE interface, the NP enrichment column was kept at 99 °C, and column regeneration time was 10 s. As shown in Fig. 1b, the switching procedure (separation case of toad venom as example) of valve was as follows: Valve A was switched every 220 s; Valve B was remained in position A for 210 s and in position B for 10 s, respectively. At elevated temperature and vacuum condition (pressure of 100 kPa), the flow of the first NPLC effluent into the NP enrichment experienced rapid solvent evaporation, which rendered fast adsorption of solute in the NP enrichment column. With the switching of Valve A, the effluent from the NPLC continued to flow into the other silica gel enrichment column. Simultaneously, Valve B was kept at position A and the second dimensional NP mobile phase was delivered to the previous column to transfer the solute for the further NPLC separation. Once the second dimensional NPLC separation completed, Valve B was switched to another position, to establish an association between the NP enrichment column and vacuum pump for column regeneration (10 s). The separations were carried out according to the conditions shown in Table 1.

On-Line 2D RPLC × RPLC Separation
Test solution of the sample EA extract of dammar resin was injected in separation with of sample loading 20 µL.

Coupling an NPLC with Another NPLC
In the past decades, various 2D-LC separation systems has been developed. In most of these systems, RP chromatography is used as second dimension, on account of high column efficiency and resolution [2]. Each of these RPLCbased combination has its application scope. NP × RP [11] and RP × RP [10] systems are commonly used to separate medium and strong polar compounds, while IEX × RP [12], HILIC × RP [13], and SEC × RP [4] systems are suitable to separate strong polar compounds. The use of RPLC as the second dimension is mainly common as a desalting step to efficiently couple all those non-MS compatible modes (SEC, HIC, IEX for example) which are now heavily used for biopharmaceuticals [14,15]. According to research, it has deficient in separating of weakly polar samples. In contrast, silica gel-based NPLC is commonly used to separate weakly polar samples and some specific biopharmaceuticals that are not suitable for separation in RPLC system.
The separation mechanism for the popular RPLC in a 2D system is based on the difference of partition coefficients, while NPLC is based on molecular polarity. For weakly polar samples that are easily adsorbed and difficult to elute, the common NPLC × RPLC and RPLC × RPLC systems are not very applicable, while NPLC is easy to elute weakly polar samples and can support wide range of solutes application, undoubtedly, an ideal choice for the separation of weakly polar components with similar polarities.

Choiceof the SE Interface
In coupling of two-dimensional RPLC, two main different strategies, sample loop or enrichment column with solvent dilution, are always applied. However, to integrate twodimensional NPLC in this study, both technical routes are no longer applicable. Unlike RPLC, which usually uses MeOH/H 2 O or ACN/H 2 O as the mobile phase, NPLC uses a more diverse and flexible solvent system. The sample loop interface cannot eliminate the influence of the solvent effect, which leads to the loss of separation efficiency in the second dimensional chromatography. When using enrichment column in 2D RPLC × RPLC system, H 2 O is generally used as a diluent to reduce the solvent strength of the first dimensional eluent. However, in NPLC, it is impossible to find a diluent that is as cheap and safe as H 2 O and can be used in a large quantity. In this study, an optimized SE interface was proposed to evaporate the NP solvent from the first dimensional NPLC in the NP enrichment column, referring to the previous reported TEAA interface in NPLC × RPLC system [16].

Optimization of the Working Condition in SE Interface
Although the SE interface designed with reference to the TEEA interface, there are still some parameters have to be investigated to ensure that it works in NPLC. A series of reference compound solutions dissolving in different NP mobile phases were transferred to the SE interface in different flow rates under different NPLC enrichment conditions. Their NPLC chromatograms (detailed in Supporting Information Figure S1 ~ S4) in the second dimensional NPLC were compared to evaluate the solvent evaporation efficiency and the sample enrichment rate.

Temperature
As the temperature of the NP enrichment column increased, the sample enrichment rate had been greatly improved, as showed in Fig. 2b. At room temperature (30 °C), the organic solvents still cannot fully evaporate even at a rather flow rate of 0.01 mL/min, and its solvent effect lead to failure to detect the sample chromatographic peak in NPLC. However, even Fig. 2 Optimization of the solvent evaporation (SE) interface. a Structure and the HPLC chromatogram of the reference compounds: 4-nitroacetophenone (1), 3,4-dimethoxy benzaldehyde (2), 4-methoxyphenol (3, 5-bromoindole (4), vanillin (5), 4-hydroxybenzaldehyde (6), 3,4-dihydroxybenzaldehyde (7), and resveratrol (2), which were mixed and solved in MeOH with each concentration of 1 mg. mL −1 , respectively. Then, the mother liquor was diluted at ratio of 1:10 by CHCl 3 (b, c) and other solvents (d); b-d influence of different parameters on the solvent evaporation efficiency and the sample enrichment rate, recorded by the NPLC chromatograms in the second dimension; b temperature; c vacuum condition; d mobile phase from the first dimension; *chromatograms marked in red refer to the optimal conditions selected for further study ◂ if the flow rate at 0.15 mL/min, the second dimension can still obtain a higher enrichment efficiency NPLC analysis spectrum when the temperature was increased to 90 °C. It can be deduced that temperature was an important factor in the working of the solvent evaporation interface.

Vacuum Condition
Under 90 °C, the NP solvent could be completely evaporated with a flow rate of 0.15 mL/min when vacuum was provided or not ( Figure S2 in the Supporting Information). However, the enrichment efficiency in a low flow rate of 0.05 mL/min was lower than that of the higher flow rate, and the effect was more obvious under vacuum condition. In general, under high-temperature conditions, the effect of vacuum on the volatilization of solvents was much lower than the effect of temperature on it.

The First Dimensional Mobile phase
For the reason that the same stationary phase was used, the orthogonality of NPLC × NPLC mainly came from by the combination of different mobile phases. The influence of different mobile phase on the SE interface was estimated (Fig. 2d).
The sample enrichment rate, transporting by CHCl 3 -MeOH (5: 1) or n-hexane-anhydrous EtOH (3: 1) at 0.20 mL/min, was preferable to the other groups. All of the NP solvents were completely evaporated, when the flow rate reached to 0.15 mL/min. However, the sample enrichment effect and chromatographic peaks were not ideal under the flow rate at 0.25 mL/min. Overall, the composition of the mobile phases had slight effect on solvent evaporation and enrichment efficiency under the same temperature and vacuum conditions.

Regeneration of Enrichment Column
Since the residual solvent will affect the efficiency of the next enrichment, the regeneration process was necessary to ensure the repeatability of the enrichment column. As previously report in TEAA interface, to obtain a rapid regeneration time, vacuum was applied and the NP enrichment column was regeneration for 10 s. By switching valve B-to-B position, the NP solvents remaining on the enrichment column volatilized and removed by high temperature and under vacuum, thereby completing the regeneration process of the enrichment column.
In summary, the optimized condition for NP enrichment was elevated temperature at 90 °C (or higher) with vacuum (Fig. 2b). Then, the enrichment column was eluted with NP solvent without restrictor in a flow rate range of 0.10-0.20 mL/min, and regenerated for 10 s with the vacuum (detailed in Supporting Information Figure S2).

On-Line NPLC × NPLC Separation of Weakly Polar Samples
The optimized condition was applied in analysis of different natural-sourced weakly polar samples. NPLC × NPLC separation on the extracts from toad venom, dammar resin, and propolis were presented.

Separation of Toad Venom
Toad venom, dried toad secretions from either B. bufo gargarizans or B. melanostictus, have been used in the traditional Chinese medicine (TCM) for thousands of years for the treatment of various diseases, including cancer, arrhythmia, and heart diseases [17,18]. Bufogenins, high liposolubility, are the major bioactive substances in toad venom, exhibiting a range of pharmacologically activities [19]. The 2D-LC separation and analysis of bufogenins in toad venom and toad skin have been reported before, using high-speed counter-current chromatography (CCC) coupled with high-performance liquid chromatography (HPLC) [20], and NPLC × RPLC [21]. Two different NPLC × NPLC separation strategies were performed in the separation of toad venom. The 2D NPLC combination of CHCl 3 /acetone and n-hexane/EtOH (Fig. 3a) achieved better separation efficiency than the combination of CHCl 3 /MeOH and n-hexane/ EtOH (Fig. 3b). The improvement was caused by the different property of acetone and MeOH, one of which is a proton acceptor solvent, and the other is a proton donor solvent.

Separation of Dammar Resin
Resin is defined as a plant secretion that hardens on exposure to air. Only a few trees can produce resin that forms amber. The genus Agathis in the family Araucariaceae are thought to produce amber type resin [22]. The amber resin richly produced by Agathis dammara (Lamb.) Rich is very famous and locally called 'dammar resin' in Indonesia and Malaysia. Dammar resin is a complex mixture composed primarily of monoterpenoid and diterpenoid compounds [22][23][24]. However, there are very few reports on the analysis of dammar resin, although it is widely used in industry and medicine [25]. Herein, the novel NPLC × NPLC method was applied in the analysis of the weakly polar resin terpenoid components. And a comparison with RPLC × RPLC separation was made.
Owing to the diversity and flexibility of normal phase solvent combinations, the novel NPLC × NPLC method (Fig. 3c) exhibited better separation orthogonality, when compared with the widely used RPLC × RPLC (Fig. 3d).

Separation of Propolis
Propolis, a resinous material produced by honeybees, has been widely used in TCM [26] and is beneficial to a variety of body systems [27]. Propolis has been analyzed mainly by RPLC [28], and has not yet been analyzed by NPLC.
In the 2D NPLC separation of propolis, thanks to the low back-pressure of NPLC, longer column could be used on-line in the second dimension. As it is shown, when a 150 mm NPLC column was used in the second dimension (Fig. 3e), excellent separation efficiency achieved, when compared with that of 50 mm column (Fig. 3f).

Conclusions
In this study, a new on-line 2D NPLC × NPLC system, integrated by a novel designed solvent evaporation (SE) interface, was established. The working parameters of the interface were evaluated comprehensively. Then, the optimized condition was applied in analysis of three different natural-sourced weakly polar samples. A better separation efficiency was achieved using the solvent combination of CHCl 3 /acetone and n-hexane/EtOH in the separation of toad venom, which gave an example of the NPLC × NPLC solvent combination optimization. In the separation of dammar resin, the novel NPLC × NPLC system exhibited better separation efficiency than that of RPLC × RPLC, owing to the diversity and flexibility of normal phase solvent combinations. In the separation of propolis, a 150 mm NPLC column was used in the second dimension to give excellent separation efficiency, thanks to the low back-pressure of NPLC. 2D NPLC × NPLC system was demonstrated to be a useful protocol for the separation of weakly polar samples.

Author Contributions
All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by LZ, BX, SL, YW, GZ, and YQ. The first draft of the manuscript was written by LZ and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding The project was supported by Fujian Provincial Health Commission Project (2017FJZYZY105).

Data availability
The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials.