Chromatographia (2013) 76:1683–1695 DOI 10.1007/s10337-013-2557-9

ORIGINAL

Characterization of Nineteen Impurities in Roxithromycin by HPLC/TOF and Ion Trap Mass Spectrometry Fan Wang • Hongxia Zeng • Jian Wang

Received: 15 June 2013 / Revised: 31 August 2013 / Accepted: 6 September 2013 / Published online: 18 September 2013 Ó Springer-Verlag Berlin Heidelberg 2013

Abstract Nineteen impurities in roxithromycin drug substance made in China were separated and identified by HPLC–MSn (TOF and TRAP) for the further improvement of official monographs in Pharmacopoeias. The fragmentation patterns and structural assignment of these impurities were studied. The column was Shim VP-ODS (250 9 4.6 mm, 5 lm). The mobile phase was 10 m mol L-1 ammonium acetate and 0.1 % formic acid aqueous solutionacetonitrile (62.5:37.5). In positive mode, full scan LC–MS was first performed to obtain the m/z value of the protonated molecules and formulas of all detected peaks on Agilent 6538Q TOF high resolution mass spectrometer. LC–MSMS and LC–MS-MS–MS were then carried out on the compounds of interest on AB SCIEX 4000 Q TRAPTM composite triple quadrupole/linear ion trap tandem mass spectrometer. The complete fragmentation patterns of nineteen impurities were studied and used to obtain information about the structures of these impurities. The structures of nineteen impurities in roxithromycin drug substance were deduced based on the HPLC–MSn data, in which nine impurities were novel impurities.

F. Wang College of Chemistry, Xiangtan University, Xiangtan, People’s Republic of China H. Zeng College of Pharmacy, Zhejiang University of Technology, Hangzhou, People’s Republic of China J. Wang (&) Zhejiang Institute for Food and Drug Control, Hangzhou 310004, People’s Republic of China e-mail: [email protected]

Keywords Column liquid chromatography
Quadrupole time-of-flight mass
Ion trap mass spectrometry
Roxithromycin
Impurity
Structure
Identification

Introduction Roxithromycin is a semi-synthetic 14-membered macrolide antibiotic in which the 9-keto group of the erythromycin A is substituted by an ether-oxime chain. Figure 1 shows the structure of roxithromycin. The erythromycin fermentation process is not entirely selective. Commercial samples of erythromycin A contain small quantities of erythromycins B, C, D, E, F, A N-oxide and N-demethyl erythromycin A. Erythromycins A and these impurities have the same the 9-keto groups, which are substituted by ether-oxime chains to form roxithromycin and its analogs. Since roxithromycin is obtained from erythromycin A, the impurities present will undergo the same modifications as erythromycin A and the analogs of roxithromycin from these impurities can be found in roxithromycin bulk samples. The reagent forming ether-oxime chain also is impure, and the analogs of roxithromycin from these impurities can be found in roxithromycin bulk samples. In addition, the degradation products of roxithromycin as well as the intermediates of the semi-synthesis may be present. The clinical effect of roxithromycin will be impacted because some impurities have less antibacterial activities and more toxicities than roxithromycin. The result of LD50 showed that the toxicity of impurity IV, named as decladinose roxithromycin, was two times as much as that of roxithromycin. So it is necessary to characterize and control these impurities to answer to the regulatory requests of FDA.

123

1684

F. Wang et al.

Instrumentation LC Apparatus An Agilent 1260 series liquid chromatography system (Agilent, USA) equipped with a binary pump and a UV detector was connected to an Agilent G1313A autosampler. Chromatographic separation was carried out at 30 °C using a Shim VP-ODS (250 9 4.6 mm, 5 lm). The mobile phase consisted of 10 m mol L-1 ammonium acetate and 0.1 %formic acid aqueous solution-acetonitrile (62.5:37.5). The flow rate was 0.8 mL min-1. Injection volume was 2.0 lL. Fig. 1 The structure of roxithromycin

The identifications of impurities could be performed by preparative liquid chromatography followed by spectrometry. It is evident that such a procedure is labor intensive and requires a large amount of material. The on-line combination of liquid chromatography and mass spectrometry (HPLC–MSn) has developed quickly as an identification tool. HPLC–MSn combines chromatographic strong separation with mass spectrometric strong qualitative advantages. Much structure information can be obtained through LC–MS-MS and LC–MS-MS–MS analysis, which shows great superiority to identify the impurities. Structural identifications and analysis of the impurities in macrolide antibiotics have been reported [1– 11]. The aim of this study was to separate and characterize unknown impurities in roxithromycin by HPLC–MSn (TOF and TRAP) for the further improvement of official monographs in Pharmacopoeias. Nineteen impurities in roxithromycin drug substance were separated and characterized, in which nine impurities were novel impurities and ten impurities have been listed in British Pharmacopoeia 2012 [12]. A complete structure was proposed for them.

Mass Spectrometry LC–MS experiment was carried out on Agilent 6538Q TOF high resolution mass spectrometer (Agilent, USA) and AB SCIEX 4000 Q TRAPTM composite triple quadrupole/linear ion trap tandem mass spectrometer (Applied Biosystems, USA). The column effluent was split using a zerodead-volume ‘‘T’’ connector, with approximately half of the flow being fed to the mass spectrometer. The mass spectrometer was equipped with an ESI source. The ionization mode was positive. The interface and parameters of TOF mass spectrometer were as follows: a capillary voltage of 3.0 kv, a cone voltage of 35 v, a source temperature of 120 °C, a dissolvation temperature of 300 °C, a cone gas flow of 50 L h-1 and a dissolvation gas flow of 400 L h-1. The interface and parameters of TRAP mass spectrometer were as follows: nebulizer pressure (40 p.s.i.), dry gas pressure (40 p.s.i.), curtain gas pressure (10 p.s.i.), dry gas temperature (550 °C), spray capillary voltage (5,000 v), CAD:high, DP:70 V, EP:10 V, CE:45 V. Sample Solution Preparation Dissolve an adequate amount of roxithromycin drug substance with mobile phase to make a solution of 1.0 mg mL-1.

Experimental Results and Discussion Chemicals and Reagents Selection of Chromatographic Conditions Roxithromycin drug substance (Batch Number: 111217), decladinose roxithromycin reference substance, erythromycin A reference substance, erythromycin 9-(E)-oxime reference substance, and (9Z)-roxithromycin reference substance was provided by Zhejiang zhenyuan Pharmaceutical Co. Ltd. (shaoxing, China). Formic acid and acetonitrile were of chromatographic grade.

123

In order to permit the use of liquid chromatography–mass spectrometry analysis, acid in the mobile phase must be a volatile acid and its concentration should be as low as possible. By testing, we found that the optimized chromatographic conditions lead to an appropriate roxithromycin retention time, and roxithromycin could be

Characterization of Nineteen Impurities

1685

completely separated from intermediates, by-products, and degradation products when using 10 m mol L-1 ammonium acetate and 0.1 %formic acid aqueous solution-acetonitrile (62.5:37.5) as the mobile phase. The TIC chromatogram of LC–MS for roxithromycin drug substance is shown in Fig. 2a, and the LC chromatogram of the official analytical method in Chinese Pharmacopoeia for roxithromycin drug substance is shown in Fig. 2b. The TIC chromatogram of LC–MS is in accordance with the LC chromatogram of the official analytical method in the peak sequence of impurities. Selection of MS Conditions APCI and ESI are two commonly used LC/MS ion source, APCI is more suitable to the analysis of neutral or small polar

compounds, while ESI is more suitable to the analysis of polar, thermo-labile alkaline or acidic compounds. According to the structural characteristic of roxithromycin and its impurities, ESI was selected to analyze the impurities in roxithromycin. Compared with negative mode, the impurities in roxithromycin had higher MS response and sensitivity in positive mode. In order to obtain the proper protonated molecules and fragmentations, various interface and mass spectrometer parameters were tested. The results showed that the interface and parameters of TOF and TRAP mass spectrometer in Mass Spectrometry was optimal. Full scan LC–MS was first performed to obtain the m/z value of the protonated molecules and formulas of all detected peaks. In order to obtain as much structural information as possible, LC–MS-MS and LC–MS-MS–MS were then carried out on the compounds of interest. The complete fragmentation

(a)

VWD1 A, Wavelength=210 nm (WJ\WJ 2012-07-31 09-42-05\LH07310000002.D) mAU

(b) 10,11

8

6

4

1

67

2

12

8

3,4

2

13

9

16 17

14 15

19

18

5

0

0

5

10

15

20

Fig. 2 Chromatograms of (a) TIC of LC–MS and (b) LC of the official analytical method for roxithromycin drug substance 1 impurity I, 2 impurity II, 3 impurity III, 4 impurity IV, 5 impurity V, 6 impurity VI, 7 impurity VII, 8 impurity VIII, 9 impurity IX, 10

25

30

35

40

45

min

roxithromycin, 11 impurity X, 12 impurity XI, 13 impurity XII, 14 impurity XIII, 15 impurity XIV, 16 impurity XV, 17 impurity XVI, 18 impurity XVII, 19 impurity XVIII

123

1686 Table 1 TOF mass spectra results of eighteen impurities in roxithromycin

F. Wang et al.

tR (min)

[M?H]? (m/z)

0.13

5.40

853.5270

C41H76N2O16

852.5197

0.02

5.75

734.4697

C37H67NO13

733.4624

0.26 (III ? IV)

6.27

749.4801

C37H68N2O13

748.4728

IV

0.26 (III ? IV)

6.28

679.4389

C33H62N2O12

678.4316

V

0.02

8.17

823.5158

C40H74N2O15

822.5085

VI

0.15

8.99

837.5323

C41H76N2O15

836.5250

VII

0.20

9.57

823.5166

C40H74N2O15

822.5093

VIII

0.44

11.35

823.5167

C40H74N2O15

822.5094

Impurity

Content (%)

I II III

Mass

IX

0.02

11.76

793.5065

C39H72N2O14

792.4993

X

–

13.49

867.5425

C42H78N2O16

866.5352

XI XII

0.56 0.11

14.34 16.05

851.5118 925.5843

C41H74N2O16 C45H84N2O17

850.5046 924.5843

XIII

0.24

17.64

821.5370

C41H76N2O14

820.5298

XIV

0.13

18.60

925.5844

C45H84N2O17

924.5843

XV

0.33

19.44

821.5372

C41H76N2O14

820.5299

XVI

0.07

21.76

851.5477

C42H78N2O15

850.5404

XVII

0.10

24.19

841.4825

C40H73N2O14Cl

840.4823

XVIII

0.23

38.40

925.5843

C45H84N2O17

924.5843

patterns of nineteen impurities were studied by MSn and used to obtain information about the structure of these impurities. The quantitative results of eighteen impurities in one sample of roxithromycin made by Zhejiang zhenyuan Pharmaceutical Co. Ltd, calculated by normalization, are shown in Table 1. TOF mass spectra results of eighteen impurities in roxithromycin are shown in Table 1. Trap mass spectra results of eighteen impurities in roxithromycin are shown in Table 2. The proposed structures of nineteen impurities in roxithromycin are shown in Fig. 3. Structure Elucidation The Theory of the Isomers That Z-isomers of several oxime erythromycin antibiotics formed by cis–trans isomerization of the oxime function when dissolved into solution, and the retention time of (9Z)isomer was shorter than that of (9E)-isomer have been reported [1]. The experimental result showed that the content of (9Z)-isomer of roxithromycin increased obviously after the solution of roxithromycin dissolved in the mobile phase was kept at room temperature for 24 h. The retention time of (9Z)-isomer of roxithromycin was shorter than that of roxithromycin, indicating more polar properties. Impurity I TOF high resolution mass showed that the formula of impurity I was C41H76N2O16. The protonated molecule at

123

Formula

m/z 853.5 of impurity I fragments into product ion at m/z 695.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 695.5 showed the loss of 121 Da, identified as ether-oxime chain C4H9O3NH2, to give a product ion at m/z 574.4. Subsequent loss of two H2O conducts to the fragment at m/z 556.4 and 538.4, a further loss of 58 Da (C3H6O) leads to the ion at m/z 480.3. These experiments revealed that the difference between impurity I and roxithromycin lied in the lactone ring. Erythromycin F, an impurity in erythromycin A, was detected in amounts of up to 0.2 %, calculated by UV normalization, in commercial samples of erythromycin A. Erythromycins A and erythromycin F have the same the 9-keto groups, which are substituted by ether-oxime chains to form roxithromycin and its analog. From the above information, it is concluded that impurity I is produced when erythromycin F undergoes the same modifications as erythromycin A. Figure 3a presents the proposed structure and fragmentation pattern of impurity I. Impurity II The protonated molecule at m/z 734.5 of impurity II fragments into product ion at m/z 576.4 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 576.4 showed loss of three H2O conducts to the fragment at m/z 558.4, 540.4 and 522.4. Subsequent loss of 58 Da (C3H6O) and 56 Da (C3H4O) leads to the ion at m/z 464.3 and 408.5. The LC retention time of impurity II was the same as that of erythromycin A

Characterization of Nineteen Impurities Table 2 Trap mass spectra results of eighteen impurities in roxithromycin

Impurity

1687

tR (min)

[M?H]? (m/z)

MS/MS Fragmentation ions (m/z)

MS/MS/MS Fragmentation ions (m/z)

574.4, 556.4, 538.4, 480.3

I

5.40

853.5

695.5

II

5.75

734.5

576.4

558.4, 540.4, 522.4, 464.3, 408.5

III

6.27

749.5

591.4

558.4, 540.4, 522.4; 434.3, 416.3, 398.3

IV

6.28

679.4

V

8.17

823.5

665.5

544.5, 526.3, 508.6

VI

8.99

837.5

679.5

558.5, 540.5, 522.5

9.57 11.35

823.5 823.5

679.5 665.5

661.4, 603.4; 558.4, 540.4, 522.4, 464.4, 408.3 544.4, 526.4, 508.4; 522.4

VII VIII IX

11.76

793.5

635.5

558.4, 540.4, 522.4, 464.3, 408.3; 478.3

X

13.49

867.5

709.5

558.4, 540.4, 522.4; 464.4

XI

14.34

851.5

677.5

659.4, 601.5; 571.4, 414.3, 396.3, 378.3

XII

16.05

925.6

767.5

691.5, 615.4

XIII

17.64

821.5

663.5

506.4

XIV

18.60

925.5

679.5

558.4, 540.5

XV

19.44

821.5

663.5

645.5, 587.5; 542.4; 506.5

XVI

21.76

851.6

693.5

572.5, 554.4, 536.4; 536.4

XVII

24.19

841.5

683.4

665.4, 603.4; 558.4, 540.4, 522.4, 464.3, 446.3; 526.4

XVIII

38.40

925.5

767.5

691.5

reference substance. The structure of impurity II was deduced as erythromycin A based on MSn data. Its structure is in agreement with the impurity A listed in British Pharmacopoeia. Figure 3b presents the proposed structure and fragmentation pattern of impurity II. Impurity III The protonated molecule at m/z 749.5 of impurity III fragments into product ion at m/z 591.4 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 591.4 showed the loss of 33 Da, identified as NH2OH, to give a product ion at m/z 558.4. Subsequent loss of two H2O conducts to the fragment at m/z 540.4 and 522.4. Another fragmentation pathway of product ion at m/z 591.4 concerned the loss of 157 Da, corresponding to desosamine, conducting to the fragment at m/z 434.3. The LC retention time of impurity III was the same as that of erythromycin 9-(E)-oxime reference substance. The structure of impurity III was deduced as erythromycin 9-(E)-oxime based on MSn data. The erythromycin 9-(E)oxime was an intermediate of the semi-synthesis. Its structure is in agreement with the impurity C listed in British Pharmacopoeia. Figure 3c presents the proposed structure and fragmentation pattern of impurity III. Impurity IV Impurity IV exhibited protonated molecule [M ?H]? at m/z 679.4 in LC–MS analysis. Its molecular mass was

equivalent to the mass of the impurity B listed in British Pharmacopoeia. The LC retention time of impurity IV was the same as that of decladinose roxithromycin reference substance. It is concluded, in accordance with the information, that impurity IV was the impurity B listed in British Pharmacopoeia, named as decladinose roxithromycin. The decladinose roxithromycin was largely produced by acid/base hydrolysis. The decladinose roxithromycin was a degradation product of roxithromycin. Figure 3d presents the proposed structure and fragmentation pattern of impurity IV. Impurity V and Impurity VIII TOF high resolution mass showed that impurity V and impurity VIII had the same formula as C40H74N2O15. Impurity V and impurity VIII exhibited the same MS spectra, MS–MS spectra and MS–MS–MS spectra. So they were suggested as isomers. The protonated molecules at m/ z 823.5 fragment into product ion at m/z 665.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 665.5 showed the loss of 121 Da, identified as ether-oxime chain C4H9O3NH2, to give a product ion at m/z 544.5. Subsequent loss of two H2O conducts to the fragment at m/z 526.3 and 508.6. Another fragmentation pathway of product ion at m/z 665.5 concerned the loss of 143 Da, corresponding to N-demethyl desosamine, conducting to the fragment at m/z 522.4. It is suggested that the absence of 14 Da, corresponding to N-demethyl, in the desosamine moiety. From the above

123

1688

information, it is concluded that the structure of impurity VIII is in agreement with the impurity F listed in British Pharmacopoeia, and impurity VIII was N-demethyl roxithromycin. N-demethyl erythromycin A, an impurity in erythromycin A, was detected in amounts of up to 0.6 %, calculated by UV normalization, in commercial samples of erythromycin A. The impurity VIII is produced when N-demethyl erythromycin A undergoes the same modifications as erythromycin A. Figure 3h presents the proposed structure and fragmentation pattern of impurity VIII. It was concluded, in accordance with the theory of the isomer in a previous section, that impurity V is a (9Z)-isomer of N-demethyl roxithromycin. Figure 3e presents the proposed structure and fragmentation pattern of impurity V. Impurity VI The protonated molecule at m/z 837.5 of impurity VI fragments into product ion at m/z 679.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 679.5 showed the loss of 121 Da, identified as ether-oxime chain C4H9O3NH2, to give a product ion at m/z 558.5. Subsequent loss of two H2O conducts to the fragment at m/z 540.5 and 522.5. TOF high resolution mass showed that impurity VI and roxithromycin had same the formula and MSn spectra. The LC retention time and MSn spectra of impurity VI was the same as that of (9Z)-roxithromycin reference substance. It was concluded that impurity VI is a (9Z)-isomer of roxithromycin, the impurity D listed in British Pharmacopoeia. Figure 3f presents the proposed structure and fragmentation pattern of impurity VI. Impurity VII The protonated molecule at m/z 823.5 of impurity VII fragments into product ion at m/z 679.5 by means of the loss of 144 Da, corresponding to mycarose. The MS3 analysis of product ion at m/z 679.5 showed loss of a H2O conduct to the fragment at m/z 661.4. Subsequent loss of 58 Da, identified as CH2=CH–OCH3, leads to the ion at m/z 603.4. Another fragmentation pathway of product ion at m/z 679.5 concerned the loss of 121 Da, identified as ether-oxime chain C4H9O3NH2, to give a product ion at m/z 558.4. Subsequent loss of two H2O conducts to the fragment at m/z 540.4 and 522.4. Further loss of 58 Da (C3H6O) and 56 Da (C3H4O) leads to the ion at m/z 464.4 and 408.3. TOF high resolution mass showed that impurity VII, impurity V and impurity VIII had same formula. The MS2 spectra of impurity VII was not same as that of impurity V and impurity VIII. The protonated molecule of impurity VII fragments into product ion by means of the loss of 144 Da, corresponding

123

F. Wang et al. Fig. 3 The proposed structure and fragmentation patterns of impurity c I (a), impurity II (b), impurity III (c), impurity IV (d), impurity V (e), impurity VI (f), impurity VII (g), impurity VIII (h), impurity IX (i), impurity X (j), impurity XI (k), impurity XII (l), impurity XIII (m), impurity (n), impurity (o), impurity XVI (p), impurity XVIII (q), impurity XIX (r)

to mycarose, but that of impurity V and impurity VIII fragment into product ion by means of the loss of 158 Da, corresponding to cladinose. From the above information, it is concluded that the structure of impurity VII is in agreement with the impurity E listed in British Pharmacopoeia. Erythromycin C, an impurity in erythromycin A, was detected in amounts of up to 2.3 %, calculated by UV normalization, in commercial samples of erythromycin A. The impurity VII is produced when Erythromycin C undergoes the same modifications as erythromycin A. Figure 3g presents the proposed structure and fragmentation pattern of impurity VII. Impurity IX TOF high resolution mass showed that the formula of impurity IX is C39H72N2O14. The protonated molecule at m/z 793.5 of impurity IX fragments into product ion at m/z 635.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 635.5 showed the loss of 77 Da, identified as ether-oxime chain C2H5O2NH2, to give a product ion at m/z 558.4. Subsequent loss of two H2O conducts to the fragment at m/z 540.4 and 522.4. Further loss of 58 Da (C3H6O) and 56 Da (C3H4O) leads to the ion at m/z 464.3 and 408.3. These experiments revealed that the difference between impurity IX and roxithromycin lied in the ether-oxime chain, 44 Da (C2H4O) less than roxithromycin. From the above information, it is suggested that impurity IX produced when an impurity in reagent, forming ether-oxime chain, joins to react. Figure 3i presents the proposed structure and fragmentation pattern of impurity IX. Impurity X The impurity X is co-eluted with roxithromycin. The protonated molecule at m/z 867.5 of impurity X fragments into product ion at m/z 709.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 709.5 showed the loss of 151 Da, identified as ether-oxime chain C5H11O4NH2, to give a product ion at m/z 558.4. Subsequent loss of two H2O conducts to the fragment at m/z 540.4 and 522.4. Further loss of 58 Da (C3H6O) leads to the ion at m/z 464.4. These experiments revealed that the difference between impurity X and roxithromycin lied in the ether-oxime chain. From the above

Characterization of Nineteen Impurities

1689

(a)

(e)

(b)

(f)

(c)

(g)

(d)

(h)

123

1690

Fig. 3 continued

123

F. Wang et al.

(i)

(m)

(j)

(n)

(k)

(o)

(l)

(p)

Characterization of Nineteen Impurities

1691

modifications as erythromycin A. Figure 3k presents the proposed structure and fragmentation pattern of impurity XI. Impurity XII and Impurity XVIII

(q)

(r) Fig. 3 continued

information, it is concluded that the structure of impurity X is in agreement with the impurity G listed in British Pharmacopoeia. The impurity X produced when an impurity in reagent, forming ether-oxime chain, joins to react. Figure 3j presents the proposed structure and fragmentation pattern of impurity X. Impurity XI TOF high resolution mass showed that the formula of impurity XI is C42H74N2O16. The protonated molecule at m/z 851.5 of impurity XI fragments into product ion at m/z 677.5 by means of the loss of 174 Da, corresponding to cladinose ?O. The MS3 analysis of product ion at m/z 677.5 showed loss of a H2O conducts to the fragment at m/z 659.4. Subsequent loss of 58 Da, identified as CH2=CH– OCH3, leads to the ion at m/z 601.5. Another fragmentation pathway of product ion at m/z 677.5 concerned the loss of 106 Da, then the loss of 157 Da, corresponding to desosamine, to give a product ion at m/z 414.3. Subsequent loss of two H2O conducts to the fragment at m/z 396.3 and 378.3. Erythromycin E, an impurity in erythromycin A, was detected in amounts of up to 4.0 %, calculated by UV normalization, in commercial samples of erythromycin A. From the above information, it is concluded that impurity XI is produced when Erythromycin E undergoes the same

TOF high resolution mass showed that impurity XII and impurity XVIII had the same formula. Impurity XII and impurity XVIII exhibited the same MSn spectra. So they were suggested as isomers. The protonated molecules at m/ z 925.6 fragment into product ion at m/z 767.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 677.5 showed loss of a H2O and loss of 58 Da, identified as CH2=CH–OCH3, lead to the ion at m/z 691.5. Further loss of a H2O and loss of 58 Da, identified as CH2=CH–OCH3, leads to the ion at m/ z 615.4. These experiments revealed that the structures of impurity XII and impurity XVIII had an ether-oxime chain C4H9O3NH2 and an alkyl ether chain C4H9O3, and an OH group of desosamine replaced by an alkyl ether chain C4H9O3. From the above information, it is concluded that the structure of impurity XVIII is in agreement with the impurity I listed in British Pharmacopoeia. Figure 3Q presents the proposed structure and fragmentation pattern of impurity XVIII. It was concluded, in accordance with the theory of the isomer in a previous section, that impurity XII is a (9Z)-isomer of impurity XVIII. Figure 3l presents the proposed structure and fragmentation pattern of impurity XII. Impurity XIII and Impurity XV TOF high resolution mass showed that impurity XIII and impurity XV had the same formula. Impurity XIII and impurity XV exhibited the same MSn spectra. So they were suggested as isomers. The protonated molecules at m/z 821.5 fragment into product ion at m/z 663.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 663.5 showed loss of a H2O conducts to the fragment at m/z 645.5. Subsequent loss of 58 Da, identified as CH2=CH–OCH3, leads to the ion at m/ z 587.5. Another fragmentation pathway of product ion at m/z 663.5 concerned the loss of 121 Da, identified as etheroxime chain C4H9O3NH2, to give a product ion at m/z 542.4. Another fragmentation pathway of product ion at m/ z 663.5 concerned the loss of 157 Da, corresponding to desosamine, conducting to the fragment at m/z 506.5. From the above information, it is concluded that the structure of impurity XV is in agreement with the impurity H listed in British Pharmacopoeia. Erythromycin B, an impurity in erythromycin A, was detected in amounts of up to 1.1 %, calculated by UV normalization, in commercial samples of

123

1692

erythromycin A. The impurity XV is produced when erythromycin B undergoes the same modifications as erythromycin A. Figure 3O presents the proposed structure and fragmentation pattern of impurity XV. It was concluded, in accordance with the theory of the isomer in a previous section, that impurity XIII is a (9Z)-isomer of impurity XV. Figure 3m presents the proposed structure and fragmentation pattern of impurity XIII.

F. Wang et al.

modification has to be located at C13 of the lactone ring. It is supposed that, in the context of biosynthetic routes, the ethyl side chain was replaced by a propyl side chain during fermentation. An analogous impurity was identified earlier in erythromycin and azithromycin sample [5]. Figure 3p presents the proposed structure and fragmentation pattern of impurity XVI. Impurity XVII

Impurity XIV The protonated molecule at m/z 925.6 of impurity XIV fragments into product ion at m/z 679.5 by means of the loss of 246 Da, 88 Da more than cladinose. The experiment suggested that the structure of impurity XIV had an ether-oxime chain C4H9O3NH2 and an alkyl ether chain C4H9O3, and an OH group of cladinose replaced by an alkyl ether chain C4H9O3. The MS3 analysis of product ion at m/z 679.5 showed the loss of 121 Da, identified as etheroxime chain C4H9O3NH2, to give a product ion at m/z 558.4. Subsequent loss of a H2O conducts to the fragment at m/z 540.5. TOF high resolution mass showed that impurity XIV, impurity XII and impurity XVIII had same formula, but the MS2 spectra of impurity XIV was not same as that of impurity XII and impurity XVIII. The protonated molecule fragments into product ion by means of the loss of 246 Da, but that of impurity XII and impurity XVIII fragment into product ion by means of the loss of 158 Da. From the above information, it is concluded that the OH group of cladinose in impurity XIV replaced by an alkyl ether chain, but the OH group of desosamine in impurity XII and impurity XVIII replaced impurity XIV by an alkyl ether chain. Figure 3n presents the proposed structure and fragmentation pattern of impurity XIV.

The protonated molecule at m/z 841.5 of impurity XVII fragments into product ion at m/z 683.4 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 683.4 showed loss of a H2O conducts to the fragment at m/z 665.4. Subsequent loss of 62 Da, identified as CH2=CH–Cl, leads to the ion at m/z 603.4. Another fragmentation pathway of product ion at m/ z 683.4 concerned the loss of 125 Da, identified as etheroxime chain C3H6O2 Cl NH2, to give a product ion at m/z 558.4. Subsequent loss of two H2O conducts to the fragment at m/z 540.4 and 522.4. Further loss of 58 Da (C3H6O) leads to the ion at m/z 464.3. Another fragmentation pathway of product ion at m/z 683.4 concerned the loss of 157 Da, corresponding to desosamine, conducting to the fragment at m/z 526.4. These experiments revealed that the difference between impurity XVII and roxithromycin lied in the ether-oxime chain. From the above information, it is concluded that the structure of impurity XVII is in agreement with the impurity J listed in British Pharmacopoeia. The impurity XVII is produced when an impurity in reagent, forming ether-oxime chain, joins to react. Figure 4 presents MS, MS/MS and MS/MS/MS spectra of impurity XVII. Figure 5 presents MSn spectra, proposed structure and fragmentation pathway of impurity XVII.

Impurity XVI Impurity XIX (oxidative degradation product) The protonated molecule at m/z 851.6 of impurity XVI fragments into product ion at m/z 693.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 693.5 showed the loss of 121 Da, identified as ether-oxime chain C4H9O3NH2, to give a product ion at m/z 572.5. Subsequent loss of two H2O conducts to the fragment at m/z 554.4 and 536.4. Another fragmentation pathway of product ion at m/z 591.4 concerned the loss of 157 Da, corresponding to desosamine, conducting to the fragment at m/z 536.4. TOF high resolution mass showed that the formula of impurity XVI is C42H78N2O15. These experiments revealed that the difference between impurity XVI and roxithromycin lied in the lactone ring, the presence of an additional 14 Da, probably corresponding to a methylene group, in the lactone ring. According to the proposed fragmentation pathway, the

123

Preparation of the oxidative degradation product 20 mg of roxithromycin was introduced into a test tube. 0.1 mL of 3 % H2O2 was added and heated for 30 min at 100 °C. After 30 min, the drug treated with 3 % H2O2 was diluted with mobile phase to 20 mL. The protonated molecule at m/z 853.5 of impurity XIX fragments into product ion at m/z 695.5 by means of the loss of 158 Da, corresponding to cladinose. The MS3 analysis of product ion at m/z 695.5 showed the loss of 121 Da, identified as ether-oxime chain C4H9O3NH2, to give a product ion at m/z 574.5. Subsequent loss of two H2O conducts to the fragment at m/z 556.5 and 538.4. The impurity XIX has a modified desosamine moiety with an additional oxygen atom. It was believed that the oxidative degradation product could correspond to N-oxide

Characterization of Nineteen Impurities

1693

Fig. 4 MS (a), MS/MS (b) and MS/MS/MS (c) spectra of impurity XVII

123

1694

F. Wang et al.

Fig. 5 Proposed fragmentation pathways of impurity XVII

roxithromycin. Figure 3r presents the proposed structure and fragmentation pattern of impurity XIX.

Conclusion This study separated and characterized the impurities in roxithromycin by HPLC–MSn (TOF and TRAP). Nineteen impurities in roxithromycin drug substance were separated and characterized, in which impurity I, impurity V, impurity IX, impurity XI, impurity XII, impurity XIII, impurity XIV, impurity XVI and N-oxide roxithromycin were novel impurities.

123

References 1. Zhong DF, Tian L, Li XQ, Xu YJ (2000) Chem J Chin U 21(1):31–35 2. Chitneni SK, Govaerts C, Adams E, Schepdael AV, Hoogmartens J (2004) J Chromat A 1056:111–120 3. Pendela M, Van den Bossche L, Hoogmartens J, Schepdael AV, Adams E (2008) J Chromat A 1180:108–121 4. Hu SG, Guo YL, Lv L (2001) J Chin Mass Spect Soc 22(2):10–16 5. Debremaeker D, Visky D, Chepkwony HK, Van Schepdael A, Roets E, Hoogmartens J (2003) Rap Comm Mass Spect 17:342–350 6. Deubel A, Fandino AS, Sorgel F, Holzgrabe U (2006) J Chromat A 1136:39–47 7. Chepkwony HK, Kamau FN, Rodriguez E, Roets E, Hoogmartens J (2001) Chromatographia 54:725–729

Characterization of Nineteen Impurities 8. Miguel L, Barbas C (2003) J Pharm Biomed Anal 33:211–217 9. Deubel A, Holzgrabe U (2007) J Pharm Biomed Anal 43:493–498 10. Van den Bossche L, Lodi A, Schaar J, Shaakovc S, Zorzand M, Tranquillini ME, Overballe-Petersene C, Hoogmartensa J, Adamsa E (2010) J Pharm Biomed Anal 53:109–112

1695 11. Kamarei F, Movaghari F, Ghaffari A, Bozchalooi IS, Zamani A, Jabbari A (2011) Arab J Chem 4:1–7 12. British Pharmacopoeia 2012:1912–1914

123