Ultra performance liquid chromatography (UPLC) analysis was utilized for the first time as a methodology for monitoring poly(amidoamine) (PAMAM) dendrimer surface transformations and product quality. Results were compared to high-performance liquid chromatography (HPLC) and were found to provide a vastly improved analytical method for the characterization of dendrimer polydispersity and variance in a typical surface modification. The application of UPLC increased the average number of theoretical plates by a factor of 7 and reduced retention times of analytes by 36%, while improving the resolution capability to discriminate surface variances in dendrimers. The new UPLC procedures were used to monitor surface modification of [core: ethylenediamine]; (G = 4); dendri-poly(amidoamine)- (i.e., [EDA]; (G4); dendri-PAMAM-) to produce biotinylated dendrimer conjugates. The enhanced sensitivity and efficiency of the UPLC analyses allowed resolution of biotin substituent levels and a better characterization of the targeted dendrimer conjugates compared to traditional HPLC methodology.
1. Introduction
As a subclass of dendritic
polymers [1–3], dendrimers have attracted considerable attention
in the past two decades largely due to the unique properties offered by this
new fourth major polymer architecture [4, 5] after traditional
(1) linear, (2) cross-linked (bridged), and (3) branched type categories. Dendrimers are readily produced in a wide variety of compositions as
precise nanostrucures with excellent control over nanoscale sizes, shapes,
nanocontainer properties, and presentation of surface chemistry [6, 7]. As such, dendrimers
are appealing to scientists as nanobuilding blocks for materials [4, 5, 8] because of their
monodisperse nanoscale sizes, guest-host encapsulation properties [9], diverse surface
chemistries, and low toxicity [10]/nonimmunogenicity
features [11]. Dendrimers have
been utilized in a variety of nanoapplications including: gene transfection
vectors [12], protein/enzyme
mimicking/modeling [13, 14], adhesion [15], as
antiviral agents [16], sensors [17], MRI contrasting
agents [18, 19], decontamination
agents [20], catalysis [21], and molecular
electronics [22] to mention a
few. Perhaps one of the most important
features of dendrimers is the ability to design mathematically defined numbers
of surface terminal groups as a function of generation. These surface groups are
often critical for designing polyvalent prototypes for nanopharmaceuticals [23, 24], nanoexcipients [25], signal
amplification [26], or targeted drug delivery [27].
Such dendrimer surface modifications often create new conjugate
structure complexity. Characterization of these conjugates requires analytical methodologies, which will allow a
thorough understanding of important parameters such as molecular weight
distribution, dendritic purity, and degree of surface functionalization. High-performance liquid chromatography
(HPLC) has traditionally been utilized to separate and characterize the purity
of various poly(amidoamine) PAMAM dendrimer generations or
conjugates [28]. HPLC has also been utilized to evaluate the
polydispersity, surface heterogeneity, and solubility of multifunctionalized
dendrimers [29] and to study the
interactions between biomolecules and dendrimers [30]. These studies as well as capillary
electrophoresis (CE) [31–33] protocols have
provided important characterization information for surface modified PAMAM
dendrimers. However, to date, most of these chromatographic methods have proven inadequate as high-resolution
characterization methodologies for monitoring the functionalization of various
dendrimer surfaces. Understanding these degrees of modification is critical in designing and characterizing these
nanostructures for various applications.
We wish to report the first use of ultra performance liquid chromatography (UPLC)
to monitor and characterize terminal group (surface) functionalization of
commercially available PAMAM (G = 4) dendrimers. UPLC builds on the principles of HPLC but utilizes smaller fixed phase
particle sizes and elevated pressures, generating higher resolution separations
of complex mixtures, increased sensitivity, and faster analysis [34, 35]. UPLC has been successfully employed for
studies such as the analysis of lipids from complex biological mixtures [36] and has provided
structural insights in peptides [37]. However, to our knowledge UPLC techniques
have not been used to characterize dendrimers, polymers, or nanomaterials.
2. Experimental Methods
2.1. Materials
Dendrimer samples; [EDA]; (G = 4); dendri-PAMAM-(N)64), and [EDA];
(G = 5); dendri-PAMAM-(N)128)
from standard Dendritic Nanotechnologies, Inc. (Mt. Pleasant, MI)
production lots were used. All DI water was purified using a Milli-Q plus water purification system (Millipore, Bedford,
MA, USA). Biotinylation of PAMAM dendrimers was accomplished using Sulfo-NHS-LC-biotin (Pierce
EZ-Link Kit) [38, 39]. Briefly, dried dendrimers were resuspended
in 0.1 M phosphate buffer saline (PBS). A 3 : 1 molar ratio of biotin/PAMAM dendrimers
was then allowed to react in 0.1 M PBS for 2 hours on an orbital shaker [38, 39]. Excess,
unreacted biotin was then removed using Microcon filters (Millipore. Bedford, MA, USA). Biotinylated dendrimer conjugates were
subsequently rinsed with PBS followed by DI water. For chromatographic analysis, all dendrimer
samples were dried and redissolved in the HPLC mobile phase, which consisted of
a 90 : 10 mixture of DI water/acetonitrile (ACN) with 0.14% trifluoroacetic acid (TFA).
2.2. High-Performance Liquid Chromatography
A reversed phase HPLC system consisting of a Waters
Breeze HPLC (Waters Corporation, Milford, MA USA), equipped with a 717 plus autosampler, 2487 dual λ UV
detector, 5 m Symmetry 300 C18 column (4.6 mm 150 mm), and a Waters Sentry Symmetry
C18 guard column was used. The mobile phase consisted of a linear gradient beginning with 90% water and 10% ACN (each
with 0.14% TFA) at a flow rate of 1 mL/min reaching 68% water 32% ACN over 10
minutes. The injection volume was 100 L, and detection of eluted samples was performed at 214 nm. These parameters were chosen based on
optimization of previous reported HPLC methods of dendrimers [28, 40–42].
2.3. Ultra Performance Liquid Chromatography
UPLC analyses were performed with a Waters Acquity Ultra Performance LC system (Waters Corporation, Milford, MA, USA). UPLC
separation was achieved on an Acquity UPLC BEH C18 column (100 mm 2.1 mm, i.d., 1.7 m
particle size, 130 Å pore size, Waters) maintained at 40°C. The compounds were injected using a partial loop injection of 7 L. Detection was done using the UPLC photodiode array detector monitoring at 214 nm. The previous HPLC method was
scaled geometrically for the UPLC system for comparison using a conversion
software program that scales all parameters of the method to the UPLC column [43]. The mobile phase consisted of 0.02% TFA in
water (solvent A) and 0.018% TFA in ACN (solvent B). The gradient used for the UPLC method
consisted of linear gradient from 99% solvent A to 65% solvent A over 14
minutes followed by a 2-minute hold and back to initial conditions. A flow rate of 0.4 mL/min was used.
3. Results and Discussion
HPLC can be used to evaluate the purity and distribution of various dendrimer generations. Figure 1 shows the HPLC analysis of G4 amine
terminated and G4 biotinylated PAMAMs. A chromatogram of G5; amine terminated
PAMAM is also included for comparison. Amine terminated PAMAMs tend to readily adsorb to a variety of
substrates, therefore reverse ion pair chromatography was conducted using TFA [29, 44].
Figure 1 clearly illustrates an enhancement
in retention time as a function of PAMAM generation, due to the geometrically
increased surface density of the terminal amine groups and TFA ion pairs. Results for G4 and G5 are similar to those
reported previously [28]. While PAMAM dendrimer populations are
relatively monodisperse compared to traditional polymer architectures, they
still exhibit minor defect levels, which may be described as generational or
skeletal dispersity (i.e., missing arms and/or intramolecular loops) [4, 5, 30]. The polydispersity of the
PAMAM dendrimers can be estimated qualitatively by assessing the peak width at
half height () [28]. Results suggest that generational defect
levels increase with dendrimer generation (e.g., G4, PAMAM, /min = 0.39, whereas, G5; PAMAM,
/min = 0.49). These trends in persisted
regardless of changes in HPLC operating conditions. Both G4 and G5, amine terminated PAMAM
generations exhibited a small shoulder to the right of the main band, which
closely overlaps with the next higher generation peak (see Figure 1). This indicates the presence of dimers
(<10% of sample), which is in agreement with previous reports in the
literature [31, 45–47]. It
should also be noted that unlike earlier
reports [28, 30], traces of trailing lower
generation PAMAMs are not apparent with the G4, PAMAM, however they are evident
with the G5, PAMAM. The G4, biotinylated
PAMAM dendrimers, exhibited broader elution profiles (/min = 1.05)
with evidence of band overlap (see
Figure 1). This elution peak profile suggests that there
are varied degrees of surface functionalization, which remain unresolved using
HPLC. The chromatogram also indicates the presence of 6% unreacted PAMAM dendrimers for the synthesis of
G4-biotinylated conjugates.
Figure 1: HPLC chromatogram of (a) [EDA core], G = 5, amine
terminated PAMAM dendrimers, (b) [EDA core], G = 4, amine terminated PAMAM
dendrimers, and (c) [EDA core], G = 4, biotinylated PAMAM dendrimers. The concentrations of each dendrimer in
solution are 0.75 mg/mL.
Figure 2 illustrates a UPLC analysis of the same G4; amine terminated PAMAM dendrimer
and biotinylated conjugate. It appears that the main elution peak associated
with G4, amine terminated PAMAM (see Figure 2(a)), consists of a series of at
least 5 differentiated overlapping bands. The dimer peak is also present along with impurities associated with TFA
(6.26, 6.87, and 7.47 minutes) and an unidentifiable impurity at 5.66 minutes. At this time, the impurity is unknown and is
absent in the analysis of biotinylated conjugates and all HPLC analyses.
Figure 2: UPLC chromatogram of (a) [EDA core], G = 4, amine
terminated PAMAM dendrimers, (b) [EDA core], G = 4, biotinylated PAMAMs and, (c)
control blank of TFA. The concentrations
of each dendrimer in solution are 0.50 mg/mL.
The UPLC chromatogram of G4 PAMAMs, Figure 2(a), provides insight into the
polydispersity of higher generation PAMAMs. These variations (or defects) are typically associated with the
successive iterations of divergent synthesis [5, 48]. The sensitivity for chromatographically
separating structural dispersities, such as dimers and trailing generations,
decreases with higher generations of PAMAMs. Researchers have also suggested that these components along with minor
skeletal defects are likely masked inside the main chromatographic peak [28, 45, 47, 49]. Results here provide details for the first
time into the skeletal differentiation of PAMAMs. Although, UPLC relies on the same
selectivity and retention methods as HPLC, the smaller particle sizes in UPLC
enable improved limits in the chromatographic separation [35]. This is evident by the presence of multiple
bands within the main peak associated with G4 (see Figure 2(a)). Furthermore, the UPLC analysis of
biotinylated G4 PAMAMs demonstrates the advantage gained by using smaller
particle sizes and higher pressures to evaluate the substitutional diversity of
dendrimers upon surface modification (see Figure 2(b)).
A comparison of HPLC and UPLC analyses of G4 bioinylated PAMAMs is shown in Figure 3. In both analyses, a peak is present
to the left of the main band, associated with unreacted G4 PAMAM. The analyses
indicate that 6% of the G4 PAMAMs were unmodified. The focus of this work, however, is to
demonstrate the new resolution capabilities to discriminate surface variances
that are possible while monitoring the synthesis of various biotinylated-PAMAM
conjugates. The UPLC chromatogram for biotinylated G4, PAMAMs (see Figure 3(b)), demonstrates an improved
separation/resolution of biotinylated PAMAM conjugates compared to the HPLC
analysis, as seen in Figure 3(a). Apart
from the G4, PAMAM peak, Figure 3(b) exhibits 6 major peaks ( = 4.98 min,
= 5.48 min, = 5.80, = 6.11 min, = 6.30, = 6.47),
associated with varied degrees of surface functionalization. NMR analysis (not shown here) indicates −17% surface coverage. However, UPLC analysis reveals that there are
significant variations in the overall degree of surface modification. If a Gaussian distribution of surface
functionalization is assumed, the present results suggest, based on peak areas, that fractions 3, 4, and 5 contain dendrimer conjugates
with larger biotin coverage, and fractions 1 and 2 contain biotinylated conjugates with a lower degree of functionalization.
Figure 3: Comparison of (a) HPLC and (b) UPLC analyses of [EDA core]; G = 4 biotinylated PAMAMs. The solution concentrations of each
dendrimer in solution are 0.50 mg/mL (*indicates TFA contaminate).
These results illustrate the enhanced potential of UPLC to assess generational defect levels as well as degree of dendrimer surface modification
in greater detail than has been possible by HPLC analyses or other chromatographic methods such as CE [31–33, 50]. Current methods utilized
to determine the degree of functionality such as NMR and matrix assisted laser desorption/ionization-time
of flight (MALDI-TOF) only provide an estimated average number of surface functional moieties. However, if UPLC could be coupled with ESI-TOF or MALDI-TOF, analysis of the more highly resolved/separated fractions
should provide deeper insights into the species identification of surface modified dendrimer conjugates.
Furthermore, UPLC allows for increased efficiency and sensitivity in analysis. This phenomena may be explained in
accordance with the van Deemter equation [51]; wherein, smaller
particles provide not only increased efficiency, but also the ability to work
at increased linear velocity without a loss of efficiency, thus providing both
resolution and speed. A summary of the separation comparison can be found in Table 1. For example, utilizing the
scaled HPLC separation method, the retention time for G4 PAMAMs is decreased
from 6.90 minutes to 4.30 minutes, reducing the
run time required for characterizing G = 4, PAMAM dendrimers by 36%. It should be noted that this reduction in run
time could potentially be improved with further modification of the UPLC method. UPLC analysis also
produces a 7-fold increase in average number of theoretical plates, improving
the capability to distinguish dendrimer surface variances. In addition, due to the reduced band
spreading during the separation process, the analytes are more concentrated at
the point of detection. This allows for lower injection volumes and concentrations of analyte, improving the limit of
detection of by a factor of 100 (HPLC method for G4, PAMAMs, is moL
compared to moL for UPLC).
Table 1: Comparison of HPLC and UPLC analyses of [EDA core] G = 4; N and G = 4; biotinylated PAMAM dendrimers. (*Indicates
dimer, †indicates unfunctionalized (G = 4; N)).
4. Conclusions
UPLC provides a significantly more detailed analysis of PAMAM dendrimer
generational defect levels. These PAMAMs
have been widely used in diagnostics [52], drug delivery [25], and gene transfection [53, 54] types of
commercial applications. Furthermore,
this new methodology has provided more resolved elution profiles for biotin
surface functionalized dendrimers. The
findings from this present study indicate that UPLC is an improved
chromatographic method for the detection, purification, and separation of
unmodified as well as surface modified PAMAM dendrimers. This characterization
method should be invaluable for the characterization of a wide variety of
surface modified dendrimers, which are known to be critical in many emerging
commercial applications [6, 55].
Acknowledgments
This project has
been funded by the Merck Institute for Science Education, Kentucky NSF EPSCoR,
Northern Kentucky University, Center for Integrated Natural Science and
Mathematics, and the NKU Research Foundation. Infrastructure support was provided, in part, by the National Center for Research Resources Grant P20 RR16481. The authors would also like to thank Waters Corporation for continued support
of research at Northern Kentucky University.