Journal of Chromatography B, 834 (2006) 178–182

Determination of spectinomycin hydrochloride and its related

substances by HPLC–ELSD and HPLC–MSn
Jian Wang a,∗ , Xiaojun Hu b , Ying Tu a , Kunyi Ni a
a Department of Analytical Chemistry, China Pharmaceutical University, Nanjing 210009, PR China
b Medical College, Zhejiang University, Hangzhou 310006, PR China

Received 22 October 2005; accepted 23 February 2006

Available online 20 March 2006

Abstract
A new and simple high-performance liquid chromatography–evaporative light scattering detection (HPLC–ELSD) method for the determination
of spectinomycin hydrochloride and its related substances was developed. The column was Agilent SB-C18 (250 mm × 4.6 mm, 5 ␮m).The mobile
phase was 25 mM trifluoroacetic acid. The drift tube temperature was 40 ◦ C. The pressure of nebulizing gas was 3.5 bar. Good separation of
spectinomycin from main related substances could be achieved. The standard curve was rectilinear in the range of 0.07–3.8 mg/ml (r = 0.9997).
Precision was 1.0% (R.S.D.). The limit of detection was 6 ␮g/ml. The method is simple and rapid, and the results are accurate and reproducible.
The HPLC–MSn method was used to characterize the structures of impurities contained in the spectinomycin. In positive mode, impurities were
elucidated by use of electrospray ion trap mass spectrometry in the multi-stage MS full scan mode. The possible structures of impurities C and D
in spectinomycin were deduced based on the HPLC–MSn data.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Spectinomycin hydrochloride; Impurities; HPLC–ELSD; HPLC–MSn

1. Introduction lytes [3–5]. The chromatographic mobile phase is nebulized with

an inert gas and evaporated in a drift tube. The response does
Spectinomycin hydrochloride belongs to a class of com- not depend on the solute optical properties, and any compound
pounds known as aminoglycoside antibiotics. Like many other less volatile than the mobile phase can be detected. Currently,
aminoglycosides, spectinomycin hydrochloride lacks a suitable the detector response is well described and shows a double loga-
chromophore, which is necessary for UV detection. For this rithmic relationship between the signals and the concentrations
reason, the analysis of spectinomycin hydrochloride is per- of analytes. Such response allows all molecules in a sample to
formed using pre-column or post-column derivation methods give a proportional signal (which means same sensitivity). This
[1,2]. Such methods, which need sample treatments, make the principle is applied in the search of impurities in pharmaceutical
HPLC system more complicated (such as reaction coil, extra products. At the same time, the structural information of impuri-
pump, etc.) and are time-consuming. Futhermore, the disad- ties in pharmaceutical products can be obtained by HPLC–MSn
vantages of derivation methods in terms of samples include: [6–10].
introduction of non-controlled impurities, degradation products The objective of this study was to develop a rapid and simple
and the most important one, unable to detect impurities of the chromatographic method which allows a direct sample introduc-
analyte lacking of the specific functional group required for tion without any derivation treatment. HPLC–ELSD method was
derivation. first used to determine the content of spectinomycin hydrochlo-
Evaporative light scattering detection (ELSD) is considered ride and its related substances. HPLC–MSn method was used to
as a universal detection mode suitable for non-absorbing ana- characterize the structures of impurities contained in the sample.
The HPLC system demonstrates a unique synergistic approach
that uses a novel low pH stable reversed phase silica column and
∗ Corresponding author. Tel.: +86 571 8645945; fax: +86 571 86459427. a simple mobile phase that was designed not to contain any non-
E-mail address: wangjian63@mail.hz.zj.cn (J. Wang). volatile reagents, which permits the use of ELSD and MS for

1570-0232/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jchromb.2006.02.045
 J. Wang et al. / J. Chromatogr. B 834 (2006) 178–182 179

detection and characterization of spectinomycin hydrochloride was used for the spectinomycin assay. A separate external stan-
and its related substances. dard calibration curve was used for the assay of low level related
substances with calibration points ranging from 1.0 to 3.0% of
2. Experimental the sample concentration (6.0 mg/ml). Calculations of the con-
tents were based on peak areas and external standard calibration
2.1. Chemicals and reagents curve. The best linear fit of the calibration curve data points was
obtained using a logarithmic type curve.
Spectinomycin hydrochloride reference substance and drug
substance (batch numbers: 020126, 020128, 020130, 030810,
030814 and 030818) were provided by Zhejiang Jinhua Kan- 3. Results and discussion
genbei Pharmaceutical Co. Ltd. (Jinhua, China). Trifluoroacetic
acid and pentafluoropropionic acid were analytical grade. 3.1. Development of the chromatography

2.2. Instrumentation 3.1.1. Selection of the mobile phase

It is difficult to retain the aminoglycosides in the reversed-
2.2.1. LC apparatus phase mode even with purely aqueous eluents. Therefore, most
An Agilent 1100 series liquid chromatography (LC) sys- chromatographic methods were based on some forms of ion-
tem equipped with a binary pump was connected to an Agilent pair modes or basic mobile phases. Perfluoranated carboxylic
G1313A autosampler. Chromatographic separation was carried acids such as heptafluorobutyric acid [4,5] were used as ion-
out at room temperature using an Agilent SB-C18 analytical pair reagents to facilitate the retention of aminoglycosides in
column (250 mm × 4.6 mm, 5 ␮m). The mobile phase consisted the reversed-phase mode.
of 25 mM trifluoroacetic acid. The flow rate was 1.0 ml/min. In order to permit the use of evaporative light scattering detec-
Detector used was a Dikma SEDEX 75 ELSD detector. The drift tion, the mobile phases can not contain any non-volatile reagents.
tube temperature was 40 ◦ C. The pressure of nebulizing gas was The acidic mobile phase consisting of 50 mM pentafluoropro-
3.5 bar. GAIN was set at 4 and 6 for assay of spectinomycin pionic acid–methanol (60:40) was tested to separate specti-
and related substances, respectively. The injection volume was nomycin from its impurities, but the chromatogram failed to
10 ␮l. demonstrate sharp symmetrical peaks and good separation.
The acidic mobile phase consisting of 25 mM trifluoroacetic
acid was tested to separate spectinomycin from its impurities
2.2.2. Mass spectrometry
and the chromatogram showed sharp symmetrical peaks and
LC–MS experiment was carried out on an Agilent1100 ion
good separation. The mobile phases consisting of various con-
trap mass spectrometer. The column effluent was split using a
centrations of trifluoroacetic acid (12.5, 25 and 50 mM) were
zero-dead-volume “T” connector, with approximately one quar-
tested to investigate the influence on peak shape, resolution and
ter of the flow being fed to the mass spectrometer. The MSD
retention time. The results showed that as the concentrations
was equipped with an ESI source. The ionization mode was
increased, the retention time of spectinomycin increased and
positive. The interface and MSD parameters were as the fol-
the chromatograms showed sharp symmetrical peaks and good
lows: nebulizer pressure [25 psi (N2 )], dry gas [N2 (8 l/min)],
separation. If concentration of trifluoroacetic acid was lower
dry gas temperature (325 ◦ C), spray capillary voltage (3500 V),
than 25 mM, spectinomycin could not be completely separated
skimmer voltage (40 V), ion transfer capillary exit (94 V), scan
from impurities. However, high concentration (50 mM) with pH
range (100–1200 m/z), spectra average (5), ion current control
below 2.0 would do harm to the chromatographic column. There-
(on), target (30,000), dwell time (300 ms).
fore, 25 mM trifluoroacetic acid was selected as the ion-pair
reagent. The results were shown in Table 1. It was shown that
2.2.3. Software the mobile phases containing methanol, acetonitrile, tetrahydro-
All data acquired were processed by Agilent Chemstation furan, acetone, and dioxane had adverse effect on the resolution
Rev. A. 09.01 software (Agilent, Palo Alto, CA). and peak shape. Therefore, the organic solvent was not added in
the mobile phase.
2.3. Procedures

2.3.1. Sample preparation Table 1

Drug substances were dissolved in water to obtain a con- Influence of concentrations of trifluoroacetic acid on resolution and peak shape
centration level within the working range. Concentrations of
Mobile pH Plate Resolution
spectinomycin solution were 1.5 and 6.0 mg/ml for the assay of phase (mM) (spectinomycin)
spectinomycin and related substances, respectively. With With
impurity D impurity C

2.3.2. Quantitation 12.5 2.16 1231 1.41 3.83

25 2.05 1621 3.66 7.80
An external standard calibration curve with calibration points 50 1.82 2047 5.19 10.28
ranging from 80 to 120% of the assay concentration (1.5 mg/ml)
180 J. Wang et al. / J. Chromatogr. B 834 (2006) 178–182

3.1.2. Influence of retention time on peak shape of

spectinomycin
It was shown that peak shape was strongly dependent on
retention time and the increase of retention time resulted in poor
peak shape. Trifluoroacetic acid (25 mM) was selected in order to
achieve a retention time of spectinomycin close to 8 min. It was
shown that at smaller retention time spectinomycin may overlap
with impurities, while at greater retention time, peak broadening
and asymmetry would increase resulting in decrease of ELSD
response factor.

3.1.3. Influence of sample concentration and injection Fig. 1. Chromatogram of related substances in spectinomycin hydrochloride. 2:
volume on peak shape Impurity D, 4: spectinomycin, 6: impurity C.
The study showed that peak shape was strongly dependent
on sample concentration and injection volume, and the increase
under acid/base hydrolysis and oxidation conditions (retention
of sample concentration and injection volume resulted in poor
time 2.6, 4.3 and 2.9 min, respectively). Spectinomycin could
peak shape. It was optimal that for assay of spectinomycin and
be completely separated from impurities C, D and degradation
related substances, the concentrations were 1.5 and 6.0 mg/ml,
products. The chromatographs for the determination of specti-
respectively, and the injection volume was 10 ␮l.
nomycin hydrochloride and its related substances were shown
in Figs. 1 and 2.
3.2. Optimization of ELSD conditions

In this study, Dikma Technologies SEDEX 75 evaporative 3.3.2. Linearity of response

light scattering detector was used and drift tube temperature It is now well known that ELSD gives non-direct linear
recommended was 40 ◦ C. Based on this, various temperatures response. A plot of log I versus log m provides a linear response
at 40, 50 and 60 ◦ C were tested to study the influence on ELSD as a plot of the peak area versus the sample concentration in
response and signal-to-noise ratio. The results showed that there double logarithmic coordinates.
was higher ELSD response at lower temperature and 40 ◦ C was For the assay of spectinomycin hydrochloride, the linearity of
optimal. response was determined by preparing in duplicate five spectino-
The pressure of nebulizing gas recommended was 3.5 bar. mycin hydrochloride solutions ranging from 50 to 250% of the
Based on this, various pressures (2.5, 3.5 and 4.0 bar) were tested assay concentration (1.5 mg/ml). Each solution was analyzed
to study the influence on ELSD response and signal-to-noise using the recommended HPLC system. The regression curve
ratio, and the results showed that 3.5 bar was optimal. was obtained by plotting log(concentration) versus log (peak
The used and recommended GAIN were 1–12 and 4–6, area). The regression equation was log A = 1.195 log C + 2.767.
respectively. Based on this, GAIN ranging from 4 to 6 was The corresponding coefficient r was 0.9997. The result indicated
tested, and the results showed that ELSD response increased good linearity.
with higher GAIN values but resulted in increase of baseline For the assay of related substances (low level linearity),
noise. When GAIN was set at 4 and 6, LOD (S/N = 3) was 9.0 five spectinomycin hydrochloride solutions were prepared with
and 6.0 ␮g/ml, respectively. Therefore, GAIN was set at 4 and 6 concentration ranging from 0.5 to 5.0% of the sample con-
for assay of spectinomycin and related substances, respectively. centration (6.0 mg/ml). The solutions were injected into the
HPLC system. The regression curve was obtained by plotting
3.3. Method validation log (concentration) versus log (peak area). The regression equa-

Preliminary method validation was performed to determine if

the HPLC system was acceptable with respect to the specificity,
linearity of response, precision and to determine the limit of
detection.

3.3.1. Specificity
The ability of the chromatographic system to resolve specti-
nomycin hydrochloride from its possible impurities was inves-
tigated. Impurities C and D were examined in order to assure
that they do not interfere (peak overlapping) with spectino-
mycin. For the related substances test, samples were stored
under relevant stress conditions (light, heat, acid/base hydrol-
ysis and oxidation, respectively). The samples showed light
and heat stability while degradation products were produced Fig. 2. Chromatogram of spectinomycin hydrochloride assay. 1: Spectinomycin.
 J. Wang et al. / J. Chromatogr. B 834 (2006) 178–182 181

tion was log A = 1.438 log C − 1.572. The corresponding coeffi-

cient r was 0.9995. The result indicated good linearity.

3.3.3. Precision of the assay

Six replicate sample solutions at 100% of the test concen-
tration (1.5 mg/ml) were prepared and then assayed for specti-
nomycin using the recommended HPLC system and sample
preparation. The relative standard deviation (R.S.D.) value was
1.0% (n = 6).

3.3.4. Limit of detection (LOD)

The limit of detection is defined as the lowest concentration of
analyte that can be accurately detected. Its determination could
be made by the calculation of the signal-to-noise ratio. A ratio
of 3 was selected and successive dilutions of the test solution
gave a LOD relative to the spectinomycin peak of 0.1% (m/m).
The limit of detection was 6.0 ␮g/ml. Such limit was in good
agreement with what required for assay of related substances. Fig. 3. (+)-ESI-MS2 spectra of impurity C.

3.4. Analysis of spectinomycin hydrochloride drug The eluent is the same as the one described above for the
substance HPLC–ELSD method.

Six batches of drug substances were analyzed using the

recommended HPLC system. The results of determination 3.5.1. The characterization of impurity C by HPLC–MSn
of spectinomycin and related substances were shown in In Fig. 1, we could find that an impurity appeared repeatedly
Tables 2 and 3. The results were in agreement with the results above an apparent level. In MS full scan mode, the major ion
of determination by the microbiological assays of antibiotics in detected was protonated molecular ion, [M + H]+ m/z 335.
Chinese Pharmacopoeia. It was demonstrated that the spectinomycin gave abundant
product ion at m/z 207 by loss of the C-ring and B-ring in
MS2 full scan mode. The possible structure of the impurity
3.5. The characterization of impurities by HPLC–MSn
in spectinomycin was deduced based on the HPLC–MSn
data (the structure is shown in Fig. 5). Its structure is in
The HPLC–MSn method was used to characterize the struc-
agreement with the impurity C listed in British Pharmacopoeia,
tures of impurities contained in the spectinomycin. The mobile
which was named as (2R,4RS,4aS,5aR,6S,7S,8R,9S,9aR,10aS)-
phase used for these experiments was described in Section 2.2.1.
2-methyl-6,8-bis(methylamino)decahydro-2H-pyrano[2,3-
b][1,4]benzodioxine-4,4a,7,9-tetrol(dihydrospectinomycins).
Table 2
Fig. 3 shows the mass spectrum of impurity C. The proposed
The results of assay determination for spectinomycin hydrochloride (n = 2)
scheme for fragmentation of impurity C in the multi-stage MS
Batches ELSD method (%) Microbiological full scan mode was shown in Fig. 5.
method (%)

020126 77.8 80.5

020128 77.3 80.9
020130 77.1 80.0
030810 77.4
030814 77.7
030818 77.2

Table 3
The results of related substances determination in spectinomycin hydrochloride
Batches Content of Content of Content of total
impurity D (%) impurity C (%) impurity (%)

020126 0.9 2.3 3.1

020128 0.8 2.1 3.3
020130 0.8 2.0 3.2
030810 0.8 1.9 2.8
030814 0.8 2.4 3.3
030818 0.5 2.5 2.6
Fig. 4. (+)-ESI-MS2 spectra of impurity D.
182 J. Wang et al. / J. Chromatogr. B 834 (2006) 178–182

on the HPLC–MSn data (the structure is shown in Fig. 6). Its

structure is in agreement with the impurity D listed in British
Pharmacopoeia, which was named as (2R,3R,4S,4aS,5aR,6S,7S,
8R, 9S, 9aR, 10aS)-2-methyl-6,8-bis(methylamino)decahydro-
2H-pyrano[2,3-b][1,4]benzodioxine-3,4,4a,7,9-pentol(dihydr-
oxyspectinomycin). Fig. 4 shows the mass spectrum of impurity
D. The proposed scheme for fragmentation of impurity D in the
multi-stage MS full scan mode was shown in Fig. 6.

Fig. 5. Chemical structure and the ESI-MS2 fragmentation pathways of impurity 4. Conclusions
C.

The described HPLC–ELSD method provides a rapid and

simple analysis for spectinomycin hydrochloride and its related
substances without derivations. The method is accurate and
reproducible. The structural information was obtained via
collision-activated dissociation and these characteristics are
applicable to the structural elucidation of impurities contained
in the spectinomycin.

References
Fig. 6. Chemical structure and the ESI-MS2 fragmentation pathways of impurity
D. [1] D.A. Stead, J. Chromatogr. B 747 (2000) 69.
[2] P. Edder, A. Cominoli, C. Corvi, J. Chromatogr. A 830 (1999) 345.
[3] R. Vogel, K. Defillipo, V. Reif, J. Pharm. Biomed. Anal. 24 (2001) 405.
3.5.2. The characterization of impurity D by HPLC–MSn [4] N.C. Megoulas, M.A. Koupparis, J. Chromatogr. A 1057 (2004) 125.
In Fig. 1, we could find that an impurity appeared repeatedly [5] I. Clarot, P. Chaimbault, F. Hasdenteufel, P. Netter, A. Nicolas, J. Chro-
above an apparent level. In MS full scan mode, the major ion matogr. A 1031 (2004) 281.
detected was protonated molecular ion, [M + H]+ m/z 351. [6] M.C. Carson, D.N. Heller, J. Chromatogr. B 718 (1998) 95.
[7] R.E. Hornish, J.R. Wiest, J. Chromatogr. A 812 (1998) 123.
It was demonstrated that the spectinomycin gave abundant
[8] M. Zhou, G. Wei, J. Chromatogr. B 798 (2003) 43.
product ion at m/z 333 by loss of the H2 O, and at m/z 207 by loss [9] A.E. Graham, E. Speicher, B. Williamson, J. Pharm. Biomed. Anal. 15
of the C-ring and B-ring in MS2 full scan mode. The possible (1997) 537.
structure of the impurity in spectinomycin was deduced based [10] M. Zhou, D. Zhong, J. Chin. Acta Pharm. Sin. 39 (1997) 826.