Structural elucidation of biologically active neomycin N-octyl derivatives in a regioisomeric mixture by means of liquid chromatography/ion trap time-of-flight mass spectrometry

Article abstract of DOI:10.1002/rcm.4534

Structural elucidation of six regioisomers of mono-N-octyl derivatized neomycin is achieved using MSn (up to n1/4 4) on an ion trap time-of-flight (IT-TOF) instrument equipped with electrospray ionization. The mixture of six derivatized neomycin analogues was generated by reductive amination in a shotgun synthetic approach. In parallel to the liquid chromatography/mass spectrometry (LC/ MS) detection, the antibacterial activity of the neomycin regioisomers was tested by post-column addition of buffer and bacterial inocula, subsequent microfractionation of the resulting mixture, incubation, and finally a chemiluminescence-based bioactivity measurement based on the pro-duction of bacterial ATP. The MS-based high-resolution screening approach described can be applied in medicinal chemistry to help in designing and producing new antibiotic substances, which is particularly challenging due to the high functionality of most antibiotic substances, therefore requiring advanced (hyphenated) separation and detection techniques for compound mixtures.

Full text of DOI:10.1002/rcm.4534

RAPID COMMUNICATIONS IN MASS SPECTROMETRY
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/rcm.4534
Structural elucidation of biologically active neomycin
N-octyl derivatives in a regioisomeric mixture by
means of liquid chromatography/ion trap time-of-flight
mass spectrometry
*
Martin Giera , Jon S. B. de Vlieger, Henk Lingeman, Hubertus Irth
and Wilfried M. A. Niessen
BioMolecular Analysis Group, Department of Chemistry, Faculty of Sciences, VU University Amsterdam, De Boelelaan 1083, 1081 HV
Amsterdam, The Netherlands
Received 30 January 2010; Revised 3 March 2010; Accepted 3 March 2010
Structural elucidation of six regioisomers of mono-N-octyl derivatized neomycin is achieved using
MSn (up to n ¼ 4) on an ion trap time-of-flight (IT-TOF) instrument equipped with electrospray
ionization. The mixture of six derivatized neomycin analogues was generated by reductive amination
in a shotgun synthetic approach. In parallel to the liquid chromatography/mass spectrometry (LC/
MS) detection, the antibacterial activity of the neomycin regioisomers was tested by post-column
addition of buffer and bacterial inocula, subsequent microfractionation of the resulting mixture,
incubation, and finally a chemiluminescence-based bioactivity measurement based on the pro-
duction of bacterial ATP. The MS-based high-resolution screening approach described can be
applied in medicinal chemistry to help in designing and producing new antibiotic substances,
which is particularly challenging due to the high functionality of most antibiotic substances,
therefore requiring advanced (hyphenated) separation and detection techniques for compound
mixtures. Copyright # 2010 John Wiley & Sons, Ltd.
In the past few years, we have been developing hyphenated
and mass spectrometry (MS)-based analytical technologies
enabling high-resolution biological activity screening of
(complex) mixtures and simultaneous identification of active
sample constituents.1–3 Such a hyphenated continuous-flow
system usually consists of an initial chromatographic
separation step, followed by the on-line addition of all
necessary biochemical reagents. MS can then be used to
monitor either the column effluent (analyte detection) or
additionally the biochemical reaction. One of the disadvan-
tages of such a continuous-flow approach is that it is only
applicable to relatively fast biological reactions. Addition-
ally, reaction constituents may hinder efficient MS detection.
In order to keep the benefits of high-resolution screening
assays, but enabling longer reaction times in the bioassay, we
have recently described a microfractionation system.4 Such
an approach has the additional advantage that the bioactivity
screening and the mass spectral identification step are
essentially decoupled (parallel assay), which allows the use
of chemicals in the bioassay which are not at all amenable
to MS detection. This is especially true if the bioactivity
screening would involve bacteria in order to screen for new
antibacterial agents. The screening for antibacterial substances
is usually carried out by incubating a small number of colony-
forming units (CFU) in the presence or absence of test
substances in a suitable broth medium.5 After incubating the
bacterial inocula under suitable conditions for 18h or longer,
the growth is monitored visually, or by the use of a UV
spectrophotometer. The concentration of a substance which
just does not lead to a detectable bacterial growth is the so-
called minimal inhibitory concentration, or MIC. Another
more sensitive readout possibility for MIC determinations is
the detection of ATP, which is formed by metabolically active
bacteria and is directly proportional to the number of bacteria
found under standardized conditions.6
In the ongoing struggle to find new antibiotics, complex
mixtures like natural extracts are still the most prominent
source for antibiotic substances7,8 or lead structures which
are synthetically modified.9 The high numbers of functional
groups and stereocenters of most antibiotics make it very
difficult to synthetically address selected groups of an
antibiotic scaffold without rebuilding the backbone with
modified building blocks. These and other difficulties in
producing novel antibiotic derivatives make it necessary to
find new strategies where separation sciences, MS, micro-
biology, and medicinal chemistry are combined in a single
platform to overcome bottlenecks of today’s drug discovery
processes.10
*Correspondence to: M. Giera, BioMolecular Analysis Group,
Department of Chemistry, Faculty of Sciences, VU University
Amsterdam, De Boelelaan 1083, 1081 HV Amsterdam, The
Netherlands.
The model compound used in this study is the aminogly-
coside antibiotic neomycin, which is an RNA targeting
substance.11 Recently, it has been shown that neomycin lipid
E-mail: mgiera@few.vu.nl
Copyright # 2010 John Wiley & Sons, Ltd.

1440 M. Giera et al.
conjugates can show potent antibacterial activity.12 More-
over, a N-per-arginylated neomycin derivative was shown to
be an antagonist of HIV Tat protein.13 Both recent develop-
ments inspired us to study the reductive amination14,15
of neomycin in a one-pot shotgun approach towards
N-alkylated derivatives and to provide a new strategy to
elucidate the bioactivity and structures of these closely
related substances generated in this way.
diluted (1:20) with LC solvent A and used without any
further cleaning for bioactivity screening.
Instrumentation and procedures
The general setup of the fully integrated MS-based high-
resolution screening system is shown in Fig. 1. The system
consists of four units: an LC separation system, an on-line MS
detection device, a microfractionator, and a plate reader,
where the bioactivity measurement after incubation is
performed.
The general setup of the fully integrated high-resolution
screening system is described here. It comprises of liquid
chromatography for the separation of complex (reaction)
mixtures, a split towards MS detection and identification
on one hand and the microfractionation system towards
bioactivity assessment on the other. The bioactivity of the
various mono-N-alkylated neomycin derivatives is com-
pared. The structural elucidation of all neomycin derivatives
was accomplished by liquid chromatography/ion trap time-
of-flight (LC/IT-TOF) MSn experiments and is discussed
in detail.
LC separation
A LC20 AB quaternary pump (Shimadzu, s’Hertogenbosch,
The Netherlands) was used to deliver a mixture of water/
acetonitrile (ACN) at 60 mL/min (solvent A: water/ACN
95:5, 0.1% acetic acid, 0.02% trifluoroacetic acid (TFA), and
solvent B: ACN/water 95:5, 0.1% acetic acid). The auto-
sampler was a Shimadzu SIL 20 A. For all experiments, an
injection volume of 5 mL was used. The column used was an
Ultra C18 (50 ꢀ 1 mm, 3 mm; Restek GmbH, Bad Homburg,
Germany). The column oven was held at 408C. A post-
column splitter (T-split, Valco, Schenkon, Switzerland) was
used to split the LC flow in a 1:2 ratio (MS:microfractionator),
that is 20 mL/min to the MS and 40 mL/min to the bioassay.
EXPERIMENTAL
Chemicals and bacteria
Bactiter Glo reagent was purchased from Promega (Leiden,
The Netherlands). Piperacillin dry substance (Ratiopharm,
Ulm, Germany) was from a local pharmacy. All other
chemicals were purchased from Sigma Aldrich (Schnelldorf,
Germany). We used E. coli BL21 (DE3) as model organism
On-line MS detection
A Shimadzu LC20 AD was used to deliver the MS make up
flow (ACN/water 60:40, 0.1% acetic acid, 0.02% TFA) at a
flow rate of 80 mL/min leading to a total flow towards the
mass spectrometer of 100 mL/min. The mass spectrometer
maintained in Muller Hinton II broth (MHII) (33 g/L).
¨
used in the on-line bioactivity assay was
a Waters
Generation of N-alkylated neomycins
Micromass Ultima Q-TOF (Waters, Milford, MA, USA).
The mass spectrometer was operated with electrospray
ionization in positive-ion mode during all experiments. The
capillary voltage was 2.5 kV, source temperature 1208C,
desolvation temperature 3508C, and the cone voltage was
35 V. Nitrogen (99.9990%) was used at flow rates of 50 L/h
as cone gas and 400 L/h for desolvation. MS data was
acquired in the range m/z 220–1250 with a spectrum
acquisition rate of 1 spectrum/s.
Stock solutions of 10 mM neomycin trisulfate (minimum 85%
neomycin B according to the supplier) in 20 mM phosphate
buffer (pH 6.5), 125 mM octanal in MeOH, and 200 mM
NaCNBH3 in 1:1 phosphate buffer/MeOH were prepared.
The octanal stock solution was diluted 1:10 with phosphate
buffer resulting in the working solution. Volumes of 200 mL
of each solution (1.25 equiv octanal) were mixed and heated
to 608C for 2.5 h. The resulting solution was subsequently
Figure 1. Schematic overview of the fully integrated MS-based high-resolution screening system. For details,
refer to the text.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm

Structural elucidation of neomycin N-octyl derivatives 1441
Microfractionation
mode, data was acquired from m/z 100–800, with an ion
accumulation time of 10 ms. MS2 experiments were per-
formed on the ion with m/z 364.226 with an ion isolation time
of 20 ms and a frequency of 45 kHz. Both collision energy and
gas were set to 50%. With an isolation time of 20 ms, MS3
experiments were performed on the ion with m/z 567.360,
using 100% collision energy and gas and a frequency of
45 kHz. MS4 settings were 30 ms isolation time, 100%
collision energy and gas, frequencies of 52 kHz and/or
83 kHz on the ions with m/z 273.217 and 275.232.
All experiments were carried out with the Gram-negative
bacterium E. coli (strain BL21 (DE3)) as model organism.
Bacterial inoculum (1:500 dilution of an overnight culture in
MHII medium, 66 g/L, about 1.1 ꢀ 106 cells/mL, according
to a 0.5 Mc-Farland standard5) and make-up buffer solution
(Tris 20 mM, pH 10.1) were pumped by a model 22 two-
channel syringe pump (Harvard Apparatus, South Natick,
MA, USA) at a flow rate of 40 mL/min. A model 234
autosampler (Gilson, Middleton, USA) served as fractiona-
tor. It was programmed to collect fractions of 6 s, resulting in
12 mL fractions. The software controlling the Gilson 234 was
home-made. The needle of the ’autosampler’ was replaced
by a 150 mm i.d. deactivated fused-silica capillary, of which
the last cm of the polyimide coating was removed and the
residual quartz glass was silylated using dimethyldichlor-
osilane in toluene (5%) for 3 min. The 384 low-volume well
plates (solid, white) were from Brand (Wertheim, Germany).
RESULTS AND DISCUSSION
Biological screening
The biological screening procedure (Fig. 1) is based on the
continuous-flow on-line addition of all necessary reagents
(E. coli bacterial inoculum and TRIS buffer for pH correction)
after LC separation, subsequent microfractionation into a
384-well low-volume microtitre plate (12 mL fractions), 18-h
Incubation and bioassay
incubation, and a chemiluminescence-based bioactivity
The incubations of the LC fractions with the E. coli bacteria
were carried out at 368C for 18 ꢁ 2 h under a humidified
atmosphere. After incubation, 12 mL of the Bactiter Glo
reagent were pipetted into each well using an eight-channel
multipipette. After addition of the Glo reagent, plates were
read at the Victor3 plate reader (Perkin Elmer, Waltham, MA,
USA) after an additional incubation time of 5 min. The plate
reader counted chemiluminescence for 1 s per well with the
aperture set to ’normal’. The readout was based on the
measurement of ATP produced by bacterial growth.6 The
levels of ATP are directly related to the number of
metabolically active bacteria within the incubation broth.
For the determination of minimal inhibitory concen-
trations (MIC), the system was operated under isocratic
conditions, taking care that the inhibitor substances eluted
without retention: 40% eluent B was used for piperacillin and
25% B for neomycin. All injections were carried out in
duplicate.
readout based on the measurement of ATP produced by
bacterial growth using the Bactiter Glo reagent.6 The reagent
extracts intracellular ATP and determines its amount with a
recombinant firefly luciferase, thereby producing light. The
levels of ATP are directly related to the number of
metabolically active bacteria within the incubation broth.
In assessing bioactivity of antibacterial agents, this means
that a high baseline in the bioactivity trace is observed when
no antibacterial agent is present and the bacterial growth can
occur without inhibition, whereas negative peaks indicate
the presence of bioactive compounds in the effluent stream.
To evaluate the effectiveness of different substances, it is
necessary to compare their peak heights obtained in the
bioassay, as the peak height is directly proportional to the
caused effect and, unlike the peak area, does not increase
anymore when a 100% response is reached. As only two-
thirds of the LC effluent is used in the bioassay, simultaneous
MS detection and identification of the bioactive substances in
the mixture using the other third of the LC effluent can be
achieved using a QTOF instrument.
MSn structural elucidation
The MSn experiments for structural elucidation were carried
out using a Shimadzu IT-TOF instrument equipped with an
ESI source operated in positive-ion mode. The curved
desolvation line and the heating block were set to 2008C,
the interface voltage was set to 5 kV, while a voltage of 1.7 kV
was applied for the detector. Nitrogen (99.9990%) was used
as a nebulizing gas at a flow of 1.5 L/min and a drying gas
flow of 10 L/min. For the fragmentation experiments, argon
(99.9995%) was used as collision gas. In the full spectrum
Prior to its application to the bioactivity screening of N-
alkylated neomycin derivatives, the screening platform was
characterized using piperacillin and neomycin as model
compounds. Some of the assay characteristics of the
described screening platform are shown in Table 1. The
obtained MIC values proved to be comparable to published
values. Usually, MIC values are determined by performing
serial 1:1 dilutions of the substance under investigation.16 In
many cases, the readout is still performed visually:17 the MIC
Table 1. Assay characteristics
Substance
Piperacillin
Neomycin
Substance
Neomycin
Piperacillin
Found MICab
0.87 mg/mL
1.04 mg/mL
Ref.: Strain/MICab
E. coli K-12 C600/1.56 mg/mL19
E. coli ATTC27853/4 mg/mL12
Intra-day repeatability (n ¼ 3), conc.a; RSDcd
Inter-day repeatability (n ¼ 3), conc.a, RSDcd
50 mg/mL; 17.8 (%)
45 mg/mL; 11.7 (%)
25 mg/mL; 16.2 (%)
30 mg/mL; 2.3 (%)
a The concentrations refer to the injected solutions.
b All MIC values were corrected for the dilution factor, as described elsewhere.4
c RSD: relative standard deviation.
d The peak ’heights’ obtained in the biological assay were compared.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm

1442 M. Giera et al.
value is the concentration of tested substance at which no
bacterial growth can be visually detected. Hence, MIC values
differing by a factor of 2–4 are not unusual, moreover
because these values are strongly dependent on incubation
conditions and used bacterial strain. Figure 2 shows a
comparison of the bioactivity data and the mass spectro-
metric data obtained during the MIC determination of
neomycin. It can be seen that no bacterial growth was
detected when a neomycin concentration of 10 mg/mL was
injected. This, of course, is also true for the injection of a
concentration of 25 mg/mL, as the maximum biological effect
is already reached. At the same time, it can be seen that at a
higher injected concentration the observed signals in the
biological trace become broader, due to band broadening
effects. Therefore, the peak height is corresponding to the
assay window (0–100% inhibition), as the peak area would
steadily increase with higher concentrations although a
maximal biological effect was already reached. This is why
we used the peak heights of the biological data and not the
peak areas for further comparisons. Acceptable intra-day
and inter-day repeatability was obtained with relative
standard deviation (RSD) below 20%. This indicated that
all parameters proved to be acceptable for a screening
technology. The effect of the organic modifier acetonitrile
was studied by performing a water injection and applying a
gradient from 0 to 90% eluent B in 10 min with a hold time of
3 min. Under these conditions, no changes in the biological
signal (that is in the bacterial growth) were observed.
Nevertheless, the maximum amount of eluent B was
restricted to 80%, as broth components started to precipitate
at about 90% of eluent B, as visually determined.
Generation of N-alkylated neomycin
derivatives
The one-pot shotgun approach applied to generate N-
alkylated neomycin derivatives by reductive amination14,15
results in a complex mixture of basically three groups of
products, that is unreacted neomycin, mono-N-alkylated
neomycin derivatives, and products from multiple N-
alkylation. In the present study, we focused on the mono-
N-alkylated neomycin derivatives, which are the major
reaction products under the reaction conditions used (1.25
equiv. of aldehyde, n-octanal in our case). The other two
groups of products are readily separated from the mono-N-
alkylated neomycin derivatives by LC.
In principle, neomycin gives six degrees of freedom for a
mono-N-alkylation reaction. Figure 3 exemplifies one of the
six N-octyl neomycin derivatives, generated by reductive
amination using n-octanal, representing an n-octyl chain on
one of the six nitrogen atoms, that is the amino groups on
the amino sugar rings 1, 2, and 4. Although the pKa values of
the nitrogen atoms are quite different,18 which may influence
the relative conversion efficiency, all six possible neomycin-
N-octyl derivatives were detected.
Figure 2. MIC determination of neomycin: (A) biological
response, (B) EIC, m/z 615.5. All injections were performed
in duplicate; the given concentrations refer to the injected
solutions.
Figure 3. Reductive amination of neomycin. The alkylated
derivatives have a single n-octyl chain on one of the six
nitrogen atoms, as exemplified for N-1. Ring numbers (in
the ring) and nitrogen numbers are specified. R ¼ H or n-octyl.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm

Structural elucidation of neomycin N-octyl derivatives 1443
Bioactivity screening of an N-octyl neomycin
shotgun mixture
discussion is primarily focused on a proof-of-principle of the
MS-based high-resolution screening platform and on the
structural elucidation of the 6 mono-N-octyl neomycin
derivatives.
The produced regioisomeric mixture of N-octyl neomycin
derivatives was 1:20 diluted and analyzed with the high-
resolution screening platform developed. A typical chroma-
togram obtained is shown in Fig. 4. The top chromatogram
was reconstructed from the plate reader data of the
individual 12 mL fractions after microfractionation and
incubation, whereas the bottom chromatogram is an
extracted ion chromatogram for m/z 727.5 and 364.3, being
the singly and doubly charged ions of the N-octyl neomycin
derivatives. It can be seen that the six generated regioisomers
were successfully separated and simultaneously screened for
bioactivity. At the present concentration, the substances 4, 5
and 6 showed full growth inhibition, whereas the substances
1, 2 and 3 did not exhibit any growth reduction. Further
experiments with other dilution factors of the reaction
mixture showed that especially compounds 4 and 6 show
significant growth inhibition. A more quantitative assess-
ment of the bioactivity of the various mono-N-octyl
neomycin derivatives is hampered by the fact that the
various analogues are generated in different ratios and
(relative) quantification is difficult due to the lack of a
chromophore, which would enable UV detection. Given the
great influence small structural changes may have on the
ionization efficiency in ESI, the observed MS response
cannot be applied for relative quantification. The present
High-resolution MSn identification of the
formed derivatives
The different neomycin derivatives were identified by high-
resolution MSn experiments using an IT-TOF hybrid
instrument coupled to LC. The full MS spectra of all six
mono-N-octyl neomycin derivatives described are identical,
showing doubly and triply charged ions, [M þ 2H]2þ and
[M þ 3H]3þ (see Fig. 5(A)). Note: under the described
conditions in the IT-TOF source, mainly doubly and triply
charged ions were observed, whereas in the Z-spray source
on the QTOF, mainly singly and doubly charged ions were
observed. For all derivatives, the doubly protonated
molecules [M þ 2H]2þ gave the most intense signals, and
these were therefore subjected to fragmentation studies.
Upon fragmentation, all doubly charged derivative ions
showed the loss of a terminal amino sugar ion (m/z 161.091)
leading to the complementary singly charged fragment ion
with m/z 567.362 (see Fig. 5(B)). When the octyl modification is
on ring 2 (cf. Fig. 3), the fragment ion with m/z 567.362 can be
generated by the loss of either of the isomeric amino sugar
rings 1 or 4. When alkylation occurs at ring 1, the fragment ion
can only be generated by the loss of ring 4, and vice versa.
MS3 experiments lead to the first observable differences
between the spectra. Four out of six N-octyl neomycin
derivatives (peaks 1, 3, 4 and 6 in Fig. 4) clearly showed a
fragment ion with m/z 273.217 (Fig. 6), while the MS3
experiments on the remaining two derivatives (peaks 2 and
5) generated the fragment ion with m/z 275.233.
The fragment with m/z 275.233 corresponds to the loss of
rings 1 and 3, indicating that the site of alkylation in peaks 2
and 5 is at ring 2. MS4 experiments on the fragment ion with
m/z 275.233 resulted in identical spectra for both derivatives,
as is expected since alkylation on either of the nitrogen atoms
Figure 4. Separation with MS detection and simultaneous
bioactivity screening of a 1:20 diluted neomycin shotgun
reaction mixture: (A) the bioactivity trace and (B) the
extracted-ion chromatogram for the ions with m/z
727.5 þ 364.3.
Figure 5. (A) MS1 and (B) MS2 spectrum on the ion with m/z
364 of peak 4. For details, refer to the text.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm

1444 M. Giera et al.
Figure 7 shows a fragmentation scheme for the different
observed ring losses, summarizing the fragmentations
described here. Next to the fragment ion with m/z 273.217,
the MS3 spectra of peaks 3 and 4 contained a fragment ion
with m/z 405.260 originating from the loss of ring 2, thus
indicating that the site of alkylation in peaks 3 and 4 is at ring
4 (see Fig. 6). The presence of a fragment ion with m/z 435.317
(loss of ring 3) in the MS3 spectra of peaks 1 and 6 confirms
that the alkylation is on ring 1.
Subsequent MS4 experiments on the ion with m/z 273.217
resulted in the formation of the fragment with m/z 226.183 for
two out of four derivatives (peaks 1 and 3). This fragment is
consistent with the loss of H2O and H2C¼NH, that is due to
cleavage of the CꢂC bond at the alkylated ring. This
fragment is crucial for the identification, since it is only likely
to occur from the isomer with a free ꢂCH2ꢂNH2 group,
leaving the ring structure intact. Additionally, a fragment
Figure 6. MS3 spectrum on m/z 567.362 of peak 4.
1 and 3 in ring 2 results in identical structures in the
symmetric fragment ion obtained after the loss of all three
unmodified rings. Therefore, no further discrimination
between these two compounds is possible.
Figure 7. Fragmentation scheme of mono-N-alkyl neomycin derivatives in MSn exper-
iments, showing the different ring losses and the calculated m/z values of the resulting
fragment ions. R ¼ n-octyl or H.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm

Structural elucidation of neomycin N-octyl derivatives 1445
The fragmentation of the mono-N-alkyl neomycin deriva-
tives in MS3 and MS4 allowed the identification of the
derivatives, except discrimination between peaks 2 and 5.
The results are summarized in Table 2. All reported m/z
values of fragment ions in MS3 experiments were within
3 ppm error of the exact mass, while all relative mass errors of
fragment ions in MS4 experiments were within 9 ppm.
CONCLUSIONS
A fully integrated high-resolution screening platform to
assess growth inhibition by antibacterial substances was
described. The described biochemical assay gave acceptable
characteristics to be used as a screening approach for
antibacterial activity. The obtained MIC values were
comparable to the literature values. The ability to screen
for bioactive substances in a complex mixture was shown for
the generated shotgun mixture of the N-octyl neomycin
derivatives. A core feature of the screening platform truly
is MS, as it enables sensitive detection and (possibly)
identification of active substances. In the presented case,
we showed that high-resolution MSn experiments on an IT-
TOF instrument can lead to a full structural assignment of
bioactive compounds. The assignment of the alkylated ring
in the N-octyl neomycin derivatives was accomplished by
investigating the subsequent cleavages of the glycosidic
bonds. Ultimately, we were able to specify the N-alkylation
site in rings 1 and 4 by the fact that the methylene-amino
derivatives showed a characteristic fragmentation in the MS4
experiments. The alkylation position at ring 2 was not
distinguishable due to the symmetry of the MS3 fragments.
The presented screening platform was designed for the fully
integrated screening of closely related substances such as
the N-octyl neomycin derivatives, for example in structure
activity studies of modified core structures. It has to be noted
Figure 8. MS4 spectra on m/z 273.217 of peak 3 (A) and 4
(B), with alkylation on either the methylamino group or the
’ring’ amino group.
with m/z 237.1955 was produced which corresponds to a loss
of two water molecules. Figure 8(A) shows the MS4 spectrum
and exemplifies the described MS4 fragment ion for
substance 3. Other MS4 fragment ions were observed for
peaks 4 and 6, as is for instance demonstrated for peak 4 in
Fig. 8(B).
Table 2. Identification of neomycin derivatives
Derivative
(peak no. according
to Fig. 4)
Characteristic
Alkylated ring
number
Alkylated
nitrogen atom
fragmentsa [m/z] in
Chemical
formula
MS3 (#) and MS4 (ꢃ)
Calculated m/z
1
2
3
1
2
4
2’
1 or 3
2’’’
#435.317
#273.217
ꢃ226.183
#407.276
#275.233
ꢃ258.208
#405.260
#273.217
ꢃ237.196
ꢃ226.180
#405.260
#273.217
ꢃ142.160
#407.276
#275.233
ꢃ258.206
#435.317
#273.217
ꢃ142.160
435.318
273.217
226.181
407.276
275.234
258.207
405.260
273.217
237.196
226.181
405.260
273.217
142.160
407.276
275.234
258.207
435.317
273.217
142.159
C20H43N4Oþ6
C14H29N2Oþ3
C13H24NOþ2
C19H39N2Oþ7
C14H31N2Oþ3
C14H28NOþ3
C19H37N2Oþ7
C14H29N2Oþ3
C14H25N2Oþ
C13H24NOþ3
C19H37N2Oþ7
C14H29N2Oþ3
C9H20Nþ
4
5
6
4
2
1
6’’’
1 or 3
6’
C19H39N2Oþ7
C14H31N2Oþ3
C14H28NOþ3
C
20H43N4Oþ6
C14H29N2Oþ3
C9H20Nþ
a found m/z, the relative error [ppm] for all MS3 measurements was below 3 ppm and for all MS4 measurements below 9 ppm.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm

1446 M. Giera et al.
full structural elucidation by means of MSn
that
a
4. Giera M, Heus F, Janssen L, Kool J, Lingeman H, Irth H. Anal.
Chem. 2009; 81: 5460.
experiments is much more difficult, if not impossible, when
screening for totally unknown substances, for example in
natural extracts or fungal broths tested for their antibacterial
activity. Nevertheless, the presented platform might be
helpful even in these cases for a first tracking of bioactive
components, followed by isolation and full structural
elucidation using NMR techniques.
5. Susceptibility Testing of Pathogens to Antimicrobial Agents - part
6: Determination of the minimum inhibitory concentration (MIC)
with the agar dilution method, DIN 58940-8, Beuth Verlag: 2002.
6. Hattori N, Sakakibara T, Kajiyama N, Igarashi T, Maeda M,
Murakami S. Anal. Biochem. 2003; 319: 287.
7. Walsh C. Nature 2000; 406: 775.
8. He H, Williamson RT, Shen B, Graziani EI, Yang HY,
Sakya SM, Petersen PJ, Carter GT. J. Am. Chem. Soc. 2002;
124: 9729.
9. He H, Shen B, Petersen PJ, Weiss WJ, Yang HY, Wang T-Z,
Dushin RG, Koehn FE, Carter GT. Bioorg. Med. Chem. Lett.
2004; 14: 279.
10. Koehn FE. J. Med. Chem. 2008; 51: 2613.
11. Foloppe N, Matassova N, Aboul-ela F. Drug Discov. Today
2006; 11: 1019.
12. Bera S, Zhanel GG, Schweizer F. J. Med. Chem. 2008; 51: 6160.
13. Litovchick A, Lapidot A, Eisenstein M, Kalinkovich A,
Borkow G. Biochemistry 2001; 40: 15612.
14. Borch RF, Bernstein MD, Durst HD. J. Am. Chem. Soc. 1971;
93: 2897.
Acknowledgements
The work of M. Giera was supported by the German Aca-
demic Exchange Service (DAAD). We thank the division of
Molecular Toxicology for granting access to a laminar air
flow bank and providing the E. coli strain. We thank the
division of Medicinal Chemistry for access to microtitre plate
readers.
15. Eggink M, Wijtmans M, Ekkebus R, Lingeman H, de Esch
IJP, Kool J, Niessen WMA, Irth H. Anal. Chem. 2008; 80:
9042.
REFERENCES
1. Irth H. In Mass Spectrometry in Medicinal Chemistry, (1st edn),
16. Andrews JM. J. Antimicrob. Chemother. 2001; 48: 5.
Wanner KT, Hofner G (eds). Wiley-VCH: Weinheim, 2007;
¨
17. Renard D, Perruchon J, Giera M, Muller J, Bracher F. Bioorg.
¨
chap. 5.
Med. Chem. 2009; 17: 8123.
2. de Boer AR, Letzel T, van Elswijk DA, Lingeman H, Niessen
WMA, Irth H. Anal. Chem. 2004; 76: 3155.
3. de Jong CF, Derks RJE, Bruyneel B, Niessen W, Irth H.
J. Chromatogr. A. 2006; 1112: 303.
18. Kaul M, Barbieri CM, Srinivasan AR, Pilch DS. J. Mol. Biol.
2007; 369: 142.
19. Minami Y, Komuro M, Sakawa K, Ishida N, Matsumoto K,
Oishi K. J. Antibiot. 1991; 44: 256.
Copyright # 2010 John Wiley & Sons, Ltd.
Rapid Commun. Mass Spectrom. 2010; 24: 1439–1446
DOI: 10.1002/rcm