ARTICLE pubs.acs.org/ac
Investigations of the Interactions between Synthetic Antimicrobial Polymers and Substrate-Supported Lipid Bilayers Using Sum Frequency Generation Vibrational Spectroscopy Christopher W. Avery,† Edmund F. Palermo,& Amanda McLaughlin,§ Kenichi Kuroda,*,†,&,‡ and Zhan Chen*,† †
Department of Chemistry, University of Michigan, Ann Arbor, Michigan 48109, United States Macromolecular Science and Engineering Center, University of Michigan, Ann Arbor, Michigan 48109, United States ‡ Department of Biologic and Materials Sciences, School of Dentistry, University of Michigan, Ann Arbor, Michigan 48109, United States § Simmons College, Boston, Massachusetts 02115, United States &
bS Supporting Information ABSTRACT: Sum frequency generation (SFG) vibrational spectroscopy was used to analyze interactions between solid-supported lipid bilayers acting as models for cellular membranes and several membrane-active random copolymers with different lipophilic side chains, named 0R (no group), 33Me (methyl group), 11Bz (benzyl group), and 33Bu (butyl group), according to both the identity and percentage of the side chains within the polymer. Biological tests of the minimum inhibitory concentration (MIC) and hemolytic concentration were performed. The inherent surface sensitivity of SFG allowed for independent monitoring of isotopically labeled lipid bilayer leaflets as a function of concentration to study polymer-bilayer interaction mechanisms. Concentrations at which each bilayer leaflet was disrupted were quantitatively determined for each copolymer. Spectroscopic evidence of interaction with the bilayer below the critical concentrations was observed for the 11Bz polymer. The lipophilic butyl side chain of the 33Bu polymer was found to be oriented parallel to the surface normal. This research shows that SFG is a useful analytical technique which provides unique details regarding the interaction mechanisms of these membrane-active copolymers and lipid bilayers.
T
he rise in bacterial resistance to antibiotics has been underway since antibiotics were first introduced in the 1930s. Recently, however, resistance has increased far more rapidly than previously observed. As a result, bacterial resistance has been labeled as a massive threat to global public health, and research into possible solutions (including replacement drugs) is underway.1 Antimicrobial peptides (AMPs) are a particularly interesting class of peptides that have the potential to yield many insights regarding the design of new drugs.1-6 Rather than target specific receptors or protein channels, AMPs target the cellular membrane itself, which minimizes the potential for bacterial resistance development.1-6 However, AMPs are inherently very complex and before they can serve as effective models for future drug development, we must first have a detailed understanding of their interaction mechanisms, including how such mechanisms may change between AMPs with different structures and functional side chains. Extensive research has been performed to understand such mechanisms.1-8 Rather than pursing complicated modeling of AMPs, some research has focused on mimicking the relevant structures and functions of AMPs via antimicrobial polymers, which have the potential for use as therapeutic agents.9 However, some specific effects of certain polymer components on the polymer/membrane interactions are still not well understood. Solid-supported lipid bilayers have been widely used as models for cellular membranes.10 However, studying peptide/bilayer r 2011 American Chemical Society
interactions with a single bilayer is extremely challenging because of the small amount of material involved in the interactions. Recently, sum frequency generation (SFG) vibrational spectroscopy has been shown to be an effective tool for investigation of lipid systems.11-28 SFG has been used to monitor both lipid monolayer and bilayer structures, lipid transmembrane movements (flip-flops), and lipid bilayer transition temperatures. Further, research has shown that SFG is a powerful method for investigating molecular mechanisms of peptide/bilayer interactions. For example, SFG studies on melittin have shown that melittin is able to adopt multiple orientations in the bilayer,16 giving further insight into the melittin/membrane interactions. The kinetics of peptide/bilayer interactions can also be followed by SFG.7,16 Previous work in our lab showed that synthetic antimicrobial oligomers with simple structures could function as effective mimics of AMPs.13,14 These oligomers were rationally designed to have similar broad-spectrum potency as natural AMPs. They could also differentiate between bacteria and erythrocyte membranes. These synthetic small oligomers are rigid and facially amphiphilic. Their interactions with lipid bilayers (as cellular membrane models) were studied in real time in situ using SFG. Received: October 1, 2010 Accepted: December 18, 2010 Published: January 13, 2011 1342
dx.doi.org/10.1021/ac1025804 | Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry
ARTICLE
Table 1. Characterization and Biological Activity of the Amphiphilic Random Copolymers polymer
Figure 1. Structure of the amphiphilic copolymers. In each polymer, both the identity of the R group and the value of the fraction of side chains containing the R group were altered to tune the biological activity and membrane interactions. Characterization of the copolymers is given in Table 1.
Another method for producing biologically active and simple AMP mimics is random copolymerization.9,29,30 Such a random copolymer is a flexible and sequence-random chain that contains lipophilic and cationic subunits, resulting in an irregular molecular structure with global amphiphilicity. It would likely interact with bilayers via very different mechanisms compared to above oligomers. Increased flexibility is very likely to impact both the activity and selectivity of the polymers, making characterization studies of such polymers very important to understanding their interaction mechanisms with lipid bilayers.31-41 In this article, we investigate several synthetic random copolymers shown to be membrane-active at very different concentrations to determine the impact of different functional lipophilic side chains on both membrane activity and the polymer/bilayer interaction mechanism. We chose to investigate copolymers containing cationic primary ammonium groups and various lipophilic groups, which have been studied in detail by Kuroda and coworkers previously. We selected the copolymers studied here as representative examples from an extensive library of methacrylate-based copolymers previously screened for their antimicrobial and hemolytic activities.42
’ EXPERIMENTAL SECTION The molecular formulas for the random copolymers studied here are shown in Figure 1. They were prepared by a previously described technique with minor alterations (Supporting Information). The antimicrobial activities and toxicities to red blood cells were tested using standard methods (Supporting Information). For SFG experiments, solid-supported lipid bilayers were built using the Langmuir-Blodgett-Schaefer method with CaF2 right-angle prisms as substrates.7,8,13-16 Both 1,2-dipalmitoyl-snglycero-3-phosphatidylglycerol (DPPG) and 1,2-dipalmitoylD62-sn-glycero-3-phosphatidylglycerol (we denote it as d-DPPG instead of DPPG-d62 in this paper for simplicity) were alternately used for the proximal (inner) or distal (outer) leaflet. Negatively charged phosphatidylglycerol is a major component of bacterial membranes.1 All lipids used in the research were purchased from Avanti Polar Lipids (Alabaster, AL). An extensive discussion of SFG theory, application, setup, and data analysis can be found in previously published work.43-67 A brief introduction of SFG and the experimental setup used in this research are included in the Supporting Information. In all cases, SFG spectra in the C-D stretching frequency region from the lipid bilayers and in the C-H stretching frequency region from the polymers were collected using a polarization combination of s-polarized SFG output, s-polarized visible input, and p-polarized IR input beams (ssp).
R
%R
DPa
MIC E. coli (μg/mL)
HC50(μg/mL)
0R
-
0
10.4
500
>2000
33Me 33Bu
methyl butyl
33 33
15.4 15.2
125 15.6
>2000 31.9
11Bz
benzyl
11
11.5
62.5
55.1
a
Number average degree of polymerization based on end-group analysis of the 1H NMR spectra.
’ RESULTS AND DISCUSSION Antimicrobial and Hemolytic Activities. The antimicrobial activity of the amphiphilic random copolymers was assessed as the minimum polymer concentration to inhibit the growth of E. coli in MH broth (Table 1). The cationic homopolymer (0R) showed only weak activity, with an MIC value of 500 μg/mL. Increasing the hydrophobicity of the polymers by including 33% methyl methacrylate units (33Me) led to a 4-fold enhancement of the activity with an MIC value of 125 μg/mL. Further increasing the hydrophobicity by elongation of the alkyl groups from methyl to butyl side chains (33Bu) caused a dramatic decrease in the MIC to 15.6 μg/mL. Potent activity against E. coli cells was also observed when a small fraction of benzyl methacrylate was included in the copolymer (11Bz), with an MIC value of 62.5 μg/mL. These results are in agreement with previously published data on related polymers.42 Leakage of hemoglobin from human red blood cells induced by each of the polymers was measured to determine the polymer concentrations required to damage mammalian cell membranes. The relatively hydrophilic copolymers 0R and 33Me showed no appreciable hemolytic activity up to the highest polymer concentration tested, which was 2000 μg/mL. On the other hand, the hydrophobic copolymer 33Bu and the copolymer containing aromatic groups 11Bz showed significant hemolytic activity (Table 1). These trends also agree closely with previously published data.42 SFG Studies on Polymers Interacting with a DPPG Bilayer. Spectrum a in Figure 2 displays the proximal leaflet of a d-DPPG/ DPPG bilayer before exposure to polymer 0R. Spectra 2b and 2c show the deuterated proximal leaflet of a similarly structured bilayer exposed to concentrations of 0R at 1.6 μg/mL and 3.1 μg/mL, respectively. As can been seen in Figure 2, no significant decrease of signal is observed. Similarly, the right panel in Figure 2 shows the distal deuterated leaflet of a DPPG/d-DPPG bilayer before (2d) and after exposure of 0R at 1.6 μg/mL (2e) and 3.1 μg/mL (2f). Again, no significant decrease in signal was observed. This lack of decrease in signal indicates that polymer 0R is not partitioning into the bilayer and disrupting the lipid ordering. Here we did not study the SFG C-H stretching signals of the d-DPPG/DPPG or DPPG/d-DPPG bilayers because the C-H signal may also be contributed by the copolymer. An identical study was done using the 33Me polymer, the SFG spectra for which are shown in Supporting Information. As for the 0R concentration dependence study, using the same concentrations, the 33Me polymer does not partition into the bilayer. The previously described antimicrobial and hemolytic activity testing showed that these polymers are inactive against bacteria and human cells at low concentrations, which corroborates the SFG result. 1343
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry
ARTICLE
Figure 2. SFG spectra from the deuterated inner leaflet of a d-DPPG/DPPG bilayer (left panel) and the deuterated outer leaflet of a DPPG/d-DPPG bilayer (right panel). Spectra show the C-D stretching region before (a, d) and after exposure to polymer 0R at two concentrations: (b, e) 1.6 μg/mL and (c, f) 3.1 μg/mL.
Figure 3. SFG spectra from the deuterated inner leaflet of a d-DPPG/DPPG bilayer (left panel) and the deuterated outer leaflet of a DPPG/d-DPPG bilayer (right panel). Spectra show the C-D stretching region before (a, f) and after exposure to polymer 11Bz at four concentrations: (b, g) 0.08 μg/mL; (c, h) 0.12 μg/mL; (d, i) 0.50 μg/mL; (e, j) 0.82 μg/mL.
A similar study of the effect of concentration was done using both d-DPPG/DPPG and DPPG/d-DPPG bilayers and increasing concentrations of 11Bz (Figure 3). In this case, the concentrations used ranged from 0.08 μg/mL to 0.82 μg/mL, a far lower and narrower concentration range than that shown for 0R and 33Me. However, unlike the 0R and 33Me polymers, a clear loss of signal from each leaflet was observed in the case of 11Bz. Signal
loss from the 2070 cm-1 symmetric stretch of the CD3 group should be due to the polymer partitioning into the bilayer and disrupting the rigid ordering of the system, or even displacing the lipids themselves from the substrate surface (we refer to both effects as “disrupting”). Signal from each leaflet did not drop simultaneously. This is most noticeable when comparing spectra 3b and 3g to 3a and 3f, respectively. Both bilayers were exposed 1344
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry
ARTICLE
Figure 4. SFG spectra from the deuterated inner leaflet of a d-DPPG/DPPG bilayer (left panel) and the deuterated outer leaflet of a DPPG/d-DPPG bilayer (right panel). Spectra show the C-D stretching region before (a, e) and after exposure to polymer 33Bu at three concentrations: (b, f) 0.03 μg/mL; (c, g) 0.17 μg/mL; (d,h) 0.25 μg/mL.
to the same concentration of 11Bz (0.08 μg/mL for 3b and 3g); however only the distal leaflet showed disruption and loss of signal (3g). Not until a concentration of 0.12 μg/mL was reached did proximal leaflet signal begin to decrease (3c). Even at the highest exposure concentration (0.82 μg/mL, 3e and 3j), when virtually all distal leaflet signal was lost, some proximal leaflet signal remained. This type of interaction mechanism has been observed in our lab previously, and it is indicative of a “molecular knife” style of interaction, where antimicrobials partition into the external surface of the bilayer before penetrating into the interior.13 Similar SFG experiments were done for the 33Bu polymer with both d-DPPG/DPPG and DPPG/d-DPPG bilayers (Figure 4). As with 11Bz, clear evidence of interaction with and disruption of the bilayer was observed. However, 33Bu yielded a clearly different interaction pattern. The span of concentrations where full signal loss was not yet achieved is both lower and far narrower for 33Bu, ranging from 0.03 μg/mL to 0.25 μg/mL. Further, in this case both the proximal and distal leaflets decreased in SFG signal simultaneously, indicating concurrent disruption of both bilayer leaflets. This is spectral evidence that 33Bu has a higher activity than all other polymers studied here, whereas 11Bz is more active than both 0R and 33Me, but less active than 33Bu. This general trend is also reflected in their antimicrobial activities (Table 1). This is especially interesting given the dramatic differences between the systems used for testing. It can be seen that the copolymer concentrations required to disrupt the lipid bilayers for 11Bz and 33Bu are much smaller compared to the MICs listed in Table 1. We believe that this is due to the fact that the copolymers studied here can specifically interact with DPPG lipids. While in the bacterial cell membrane, many other lipids and biological molecules exist together, thus requiring a higher copolymer concentration to disrupt the bacterial cell membrane. A similar phenomenon was observed while studying AMP-lipid bilayer interactions using SFG.66 In addition, in the case of antimicrobial activity
testing (and unlike in the SFG studies), MIC values were determined in broth. Despite these different cell membranes and experimental conditions, the same trend of the activities of various polymers is observed. There are several possible reasons for the loss of SFG signal as a function of concentration, as shown above. One, already described, is that polymers can partition into the bilayer to disrupt the bilayer ordering or even displace the bilayer from the substrate. Such disruption would strongly decrease the SFG signal. However, another possibility is transmembrane movement of the lipids across the hydrophobic core of the bilayer itself, referred to as flip-flop. Such movement would mix the bilayer lipids, decreasing the signal for each leaflet by decreasing the asymmetric separation of the isotopically labeled lipids. It has been well-documented that flip-flop is an important consideration in SFG studies, as its occurrence can significantly impact the observed results.68,69 It is our contention that under the experimental conditions described here, this does not occur. The independent signal loss from each bilayer leaflet in the case of the 11Bz polymer is strong evidence for this, as such a pattern of signal loss is inconsistent with transmembrane movement of lipids. Further, under these experimental conditions, the DPPG bilayer should be in the gel phase, making the intrinsic flip-flop rate of the bilayer so slow as to be negligible. The formation of transmembrane pores has been documented to occur when a bilayer interacts with a membraneactive peptide,66 and the formation of such pores would greatly increase the flip-flop rate, as it would provide a hydrophilic gateway for the polar lipid headgroup to pass through the hydrophobic bilayer core. It is our contention that formation of such transmembrane pores is unlikely. Polymer 11Bz shows independent action by the polymer on each of the bilayer leaflets, which is inconsistent with the expected effect of a transmembrane pore. While 33Bu shows a simultaneous loss of signal, which could be consistent with disruption, flip-flop, or a combination of both, results from our time-dependent study (shown later) 1345
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry
Figure 5. Time-dependent SFG signals monitoring the peak intensity at 2070 cm-1 for a d-DPPG/d-DPPG symmetric bilayer with a D2O subphase during exposure to 11Bz at (a) 0.50 μg/mL and (b) 0.08 μg/mL.
indicate flip-flop is not occurring. Therefore, it is most probable that 11Bz and 33Bu are interacting with and directly disrupting the bilayer, as opposed to forming pores. SFG Signals from the Polymers. To determine the orientation of the lipophilic side chains for each of the copolymers, a symmetric d-DPPG/d-DPPG bilayer was built using D2O as a subphase. There are two nondeuterated methylene groups in the headgroup region of d-DPPG, but they are in a symmetric arrangement and would unlikely generate SFG C-H stretching signals. Because the bilayer acyl chains and subphase are entirely deuterated, the polymer itself is the only source of C-H or O-H vibrational stretching signals. To ensure that any signal in the C-H stretching regions would be due to the polymer and not lipids, SFG scans were taken of d-DPPG monolayers and bilayers in the C-H region before polymer was introduced to the subphase. In all cases, no SFG C-H stretching signal from the lipids was observed. No SFG C-H stretching signal was observed after the bilayer contacted the 0R and 33Me polymer solutions, which was expected, as both polymers were not active to a DPPG bilayer, as previously discussed. Additionally, no C-H stretching signal was observed from the 11Bz polymer. This was surprising because 11Bz was shown to be bilayer active and the C-D stretching signal due to the breakdown in symmetry of the d-DPPG/d-DPPG bilayer was observed. This is a particularly interesting result, because the C-D signal observed from a fully deuterated bilayer during perturbation by the polymers is evidence that the polymers are interacting with the bilayer in a way that they unevenly disrupt the two leaflets. If both leaflets were disrupted simultaneously, the disruption would maintain relative symmetry between the leaflets, yielding no C-D stretching SFG signal. The observed increase in SFG signal here implies that one leaflet is disrupted more extensively than the other. As shown in Figure 5, prior to exposure to 11Bz, no bilayer signal was observed. However, after injection of 11Bz into the subphase at 100 s, some bilayer signal was detected. This is due to the polymer partitioning into the distal leaflet of the bilayer and disrupting it, also disrupting the symmetry of the bilayer overall, and allowing for the SFG signal to be observed. Despite this evidence of interaction, no corresponding polymer signal was observed in the C-H stretching region. Possibly it is because of an orientation effect and a
ARTICLE
number density issue. If 11Bz is not oriented along the bilayer plane in a regular manner (e.g., the benzyl rings may more or less lie down on the lipid bilayer surface), the polymer will not give a SFG signal. Further, SFG signals are dependent upon the number density of the SFG-active moieties present at a surface; the fewer SFG-active moieties present, the lower the SFG signal observed. Only 11% of the subunits contain the lipophilic benzyl ring. Because 11Bz had a degree of polymerization of roughly 12 (see Table 1), this translates to only one benzyl ring per polymer on average. It is possible that there are simply not enough SFGactive moieties present at the surface when 11Bz interacts with a lipid bilayer to produce recognizable SFG signal patterns. Besides the benzyl rings, another possible source of SFG signal in 11Bz is the R-methyl groups, located along the backbone of the polymer (see Figure 1). SFG R-methyl stretching modes can be quite strong; however, no such signals are observed in this case. Therefore, the R-methyl groups on the 11Bz polymer chain may have a random distribution, giving them inversion symmetry and resulting in no observable SFG signal. A third possible source of SFG C-H stretching signal in 11Bz is methylene stretches. Again, as no SFG signals are observed at any corresponding methylene stretches, we conclude that these moieties have inversion symmetry and are therefore SFG-inactive. Exposures of 33Bu to d-DPPG/d-DPPG bilayers with a D2O subphase were also performed at several concentrations (Figure 6). As in the case of 11Bz, clear evidence of interaction with the bilayer was observed, as the symmetry of the deuterated bilayer was broken at both concentrations, shown in the left panel of Figure 6. However, signal from the 33Bu polymer itself was also observed, shown in the right panel of Figure 6. Signals in Figures 6b and 6d were taken at a subphase concentration of 0.03 μg/mL, at which perturbation of both leaflets had begun to occur but was still minimal (see Figure 4). This was further supported by the time-dependent spectra monitoring of the 2070 cm-1 peak of the terminal CD3 group of the d-DPPG lipid molecules. Before 33Bu was injected into the subphase, no bilayer signal was observed. Upon injection (at about 150 s) some bilayer signal was detected as the polymer disrupted the distal leaflet. This signal due to a break in the bilayer symmetry increased quickly and then decreased again, because the polymer penetrated further into the bilayer and disrupted the proximal leaflet, which reduced the “asymmetry” in the two leaflets. Apparently, the increase and then the decrease of the C-D signal should not be due to the flip-flop process, because otherwise the two leaflets should behave more or less similarly. At about the 30 min timepoint shown in Figure 6b, SFG C-H stretching spectra were taken. Spectrum 6d showed both symmetric and asymmetric stretching signals due to the terminal CH3 group on the butyl lipophilic side chain. Similarly, signals in Figures 6a and 6c were taken at a subphase concentration of 0.13 μg/mL, a concentration at which significant penetration into the bilayer’s proximal leaflet had already occurred, which was determined in the previously described concentration study (Figure 4). Again, upon injection of 33Bu into the subphase, a bilayer signal was detected. Perturbation of the distal leaflet broke the symmetry of the d-DPPG/d-DPPG bilayer, and again, this signal dropped as the polymer penetrated further into the bilayer. This also indicates that the signal change is not due to flip-flop. Also of note is the similarity of results between these time-dependent studies using symmetrically deuterated lipid bilayers and the previously discussed asymmetric lipid bilayers used for concentration dependent studies. For 0.13 μg/mL, the C-D signal for the symmetric bilayer returns to zero because the 1346
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry
ARTICLE
Figure 6. (Left panel) Time-dependent SFG signals monitoring the peak intensity at 2070 cm-1 for a d-DPPG/d-DPPG symmetric bilayer with a D2O subphase during exposure to 33Bu at (a) 0.13 μg/mL and (b) 0.03 μg/mL. (Right panel) SFG spectra of 33Bu at (c) 0.13 μg/mL and (d) 0.03 μg/mL interacting with a d-DPPG/d-DPPG bilayer with a D2O subphase. The results from spectral fittings are shown.
two leaflets were finally disrupted more or less to the same extent, matching observations shown in Figure 4c and Figure 4g. By comparison, at 0.03 μg/mL polymer concentration, at 30 min, the SFG signal still decreases, perhaps because it has not yet reached the equilibrium. Figure 4b also shows a slightly stronger signal than that in Figure 4f, further indicating agreement between these experiment results and those previously described. Polymer Orientation. To deduce the orientation of the terminal methyl group, orientational analysis of the methyl symmetric (2880 cm-1 and 2940 cm-1) and asymmetric mode (2960 cm-1) of the terminal methyl group on the butyl lipophilic side chain was performed.52 It is well-known that R-methyl peaks can be observed by SFG if they are present on the surface and do not lie down, typically located around 2930 cm-1 for the symmetric stretch and 2960 cm-1 and 2990 cm-1 for asymmetric stretches.51 According to previous results,51,70 R-methyl moieties tend to be oriented randomly or semirandomly on polymer surfaces. Also, the lack of any R-methyl SFG signal from 11Bz is strong evidence that the R-methyl moieties along the backbone of these polymers are randomly oriented within the bilayer. If we assume random orientation of R-methyl moieties, we would not expect the R-methyl peak to interfere with the asymmetric stretching mode of the terminal methyl group. Under this assumption, the red spectra in Figure 6 show the fitting results for 33Bu. The details of the fitting parameters for each concentration are shown in Supporting Information. According to the measured |χyyz,as/χyyz,s|, we could deduce the orientation angle of the side chain end methyl groups.71,72 The ratio |χyyz,as/ χyyz,s| for 33Bu at each of these concentrations is nearly identical. At 0.03 μg/mL, the ratio is calculated to be 0.05, whereas at 0.13 μg/mL, the ratio is calculated to be 0.06. This corresponds to a limited range of possible orientations (θ < 20°) and orientation distribution width (σ < 10° assuming a Gaussian orientation distribution). These results point toward a slight tilt from the surface normal with a reasonably narrow orientational distribution, indicating that the butyl side chain is oriented roughly parallel to the bilayer surface normal. These results seem to indicate that 33Bu interacts with the bilayer by penetrating
into the lipid bilayer core with the acyl chains. This mechanism does not seem to be concentration dependent, given the similar orientation results at multiple concentrations spanning the range of bilayer-active subphase concentrations. Membrane Disruption. These results yield some interesting clues to the interaction mechanisms between antimicrobial polymers and lipid bilayers. First, it is clear from the comparison of the activities of polymers with differing R groups that the identity of the antimicrobial moiety has a significant impact on interaction. The polymers with 0R and 33Me groups had much higher MICs than did 33Bu or 11Bz, while the latter two showed far more hemolytic activity. Of the two polymers which were active at lower concentrations (33Bu and 11Bz), significantly different interaction mechanisms were observed. While no specific orientation was able to be determined for 11Bz, it was clearly seen that the inner and outer bilayer leaflets were independently disrupted. The lack of SFG signal from the benzyl rings perhaps indicates that they more or less lie down on the membrane surface. They may interact/ disrupt the outer leaflet first and then interact/disrupt the inner leaflet. In contrast, the two leaflets were disrupted to more or less the same extent for 33Bu. This could be partially due to the higher level of activity against the bilayer for 33Bu than for 11Bz. However, orientation analysis of the terminal methyl group for 33Bu provides additional interaction information. The terminal methyl group was found to be oriented roughly parallel to the surface normal of the bilayer. The majority of the butyl chain can thus extend into the hydrophobic core of the bilayer to directly disrupt the lipid acyl chains and cause the loss of SFG signal from the bilayer. It is particularly interesting that the butyl chains exhibit this interaction mechanism whereas the benzyl rings do not. One possible explanation for the probable lack of quick penetration into the hydrophobic core by the benzyl groups is steric hindrance. A DPPG symmetric bilayer in the gel phase is tightly packed, and a benzyl group would require a greater degree of disruption than would a butyl chain. Additionally, a butyl group (while far shorter than the 16-carbon acyl chain on a DPPG lipid) is still far more similar to the chain than a benzyl ring. It is possible that the 1347
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry principle of “like dissolves like” may apply in this instance; polymers with antimicrobial groups structurally similar to structures within the lipids could be more likely to penetrate into the core of the bilayer at lower concentrations. These results provide useful guidance in the design and development of future antimicrobial polymers. It is very clear that the identity of the lipophilic group plays a major role in determining the antimicrobial character of the polymer, both in terms of its bulk effect (referring to antimicrobial activity and hemolytic toxicity) and its molecular level interaction mechanisms. Polymers with acyl chains of at least four carbons show the best ability to interrupt the bilayer. Future studies of the kinetics of these polymers and the use of lipid bilayers with different charges and phases should provide greater insight into the detailed polymer-bilayer interaction mechanism, yielding further guidance in the design of therapeutic polymeric compounds.
’ CONCLUSIONS This paper investigates the interaction between several different membrane-active copolymers and lipid bilayers via SFG. By monitoring the C-H and C-D stretches within the bilayer and the polymer itself, we were able to characterize the interaction between the two and correlate it with the antimicrobial activity data. SFG studies indicated that polymers 0R and 33Me were membrane-inactive up to 3.1 μg/mL, whereas 11Bz and 33Bu showed potent activity in all tests. Further, we found spectroscopic confirmation of a higher level of activity for 33Bu than for 11Bz. The 11Bz polymer is largely made up of cationic subunits, with on average only one lipophilic benzyl-containing side chain on each polymer molecule. It is possible that this may be related to its somewhat lower activity. SFG spectra collected from the 11Bz and 33Bu polymers themselves yielded dramatically different results. The 11Bz polymer yielded no recognizable SFG signals, indicating either a lack of regular orientation of specific moieties along the bilayer surface or a lack of enough polymer at the surface to give rise to a significant SFG signal. In contrast, 33Bu showed clear orientational organization, with the terminal methyl group oriented approximately parallel to the surface normal of the bilayer. Taken together, these results paint an interesting picture of the potential interactions mechanisms by which antimicrobial polymers operate. It is our belief that SFG is a unique, sensitive, and powerful technique for investigating both bilayers and membrane-active biomolecules, as well as testing the efficacy and activity of synthetic membrane-active compounds such as those under study here. Fundamental studies on systems of membranes and the systems that associate with them are both important and critical to the pharmaceutical industry as it seeks to create new drugs. We hope to continue to develop SFG into a powerful bioanalytical tool. ’ ASSOCIATED CONTENT
bS
Supporting Information. More details regarding experiments and results. This information is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected] or
[email protected]. Fax: 734647-4685.
ARTICLE
’ ACKNOWLEDGMENT This research is supported by Office of Naval Research (N00014-08-1-1211) and NSF CAREER Award (DMR-0845592). ’ REFERENCES (1) Zasloff, M. Nature 2002, 415, 389–395. (2) Matsuzaki, K. Biochem. Biophys. Acta 1999, 1462, 1–10. (3) Hancock, R. E. W.; Diamond, G. Trends Microbiol. 2000, 8, 402–411. (4) Epand, R. M.; Vogel, H. J. Biochem. Biophys. Acta 1999, 1462, 11–28. (5) Sitaram, N.; Nagaraj, R. Biochem. Biophys. Acta 1999, 1462, 29–54. (6) Brogden, K. A. Nat. Rev. Microbiol. 2005, 3, 238–250. (7) Chen, X.; Chen, Z. Biochem. Biophys. Acta 2006, 1758, 1257–1273. (8) Chen, X.; Wang, J.; Kristalyn, C. B.; Chen, Z. Biophys. J. 2007, 93, 1–10. (9) Palermo, E. F.; Kuroda, K. Appl. Microbiol. Biotechnol. 2010, 87, 1605–1615. (10) Tamm, L. K.; McConnell, H. M. Biophys. J. 1985, 47, 105–113. (11) Anderson, N. A.; Richter, L. J.; Stephenson, J. C.; Briggman, K. A. J. Am. Chem. Soc. 2007, 129, 2094–2100. (12) Anglin, T. C.; Conboy, J. C. Biophys. J. 2007, 95, 186–193. (13) Chen, X.; Tang, H.; Even, M. A.; Wang, J.; Tew, G. N.; Chen, Z. J. Am. Chem. Soc. 2006, 128, 2711–2714. (14) Avery, C. W.; Som, A.; Xu, Y.; Tew, G. N.; Chen, Z. Anal. Chem. 2009, 81, 8365–8372. (15) Chen, X.; Boughton, A. P.; Tesmer, J. J. G.; Chen, Z. J. Am. Chem. Soc. 2007, 129, 12658–12659. (16) Chen, X.; Wang, J.; Boughton, A. P.; Kristalyn, C. B.; Chen, Z. J. Am. Chem. Soc. 2007, 129, 1420–1427. (17) Doyle, A. W.; Fick, J.; Himmelhaus, M.; Eck, W.; Graziani, I.; Prudovsky, I.; Grunze, M.; Maciag, T.; Neivandt, D. J. Langmuir 2004, 20, 8961–8965. (18) Harper, K. L.; Allen, H. C. Langmuir 2007, 23, 8925–8931. (19) Kim, J.; Kim, G.; Cremer, P. S. Langmuir 2001, 17, 7255–7260. (20) Levy, D.; Briggman, K. A. Langmuir 2007, 23, 7155–7161. (21) Liu, J.; Conboy, J. C. J. Am. Chem. Soc. 2004, 126, 8894–8895. (22) Liu, J.; Conboy, J. C. Biophys. J. 2005, 89, 2522–2532. (23) Ma, G.; Allen, H. C. Langmuir 2007, 23, 589–597. (24) Nickolov, Z. S.; Britt, D. W.; Miller, J. D. J. Phys. Chem. B 2006, 110, 15506–15513. (25) Sovago, M.; Wurpel, G. W. H.; Smits, M.; Muller, M.; Bonn, M. J. Am. Chem. Soc. 2007, 129, 11079–11084. (26) Watry, M. R.; Tarbuck, T. L.; Richmond, G. L. J. Phys. Chem. B 2003, 107, 512–518. (27) White, R. J.; Zhang, B.; Daniel, S.; Tang, J. M.; Ervin, E. N.; Cremer, P. S.; White, H. S. Langmuir 2006, 22, 10777–10783. (28) Oh-e, M.; Sasaki, T.; Noi, M.; Goto, Y.; Itoh, K. Anal. Bioanal. Chem. 2007, 3888, 73–79. (29) Kuroda, K.; DeGrado, W. F. J. Am. Chem. Soc. 2005, 127, 4128. (30) Palermo, E. F.; Sovadinova, I.; Kuroda, K. Biomacromolecules 2009, 10, 3098–3107. (31) Yang, L.; Gordon, V. D.; Mishra, A.; Som, A.; Purdy, K. R.; Davis, M. A.; Tew, G. N.; Wong, G. C. L. J. Am. Chem. Soc. 2007, 129, 12141–12147. (32) Yang, L.; Gordon, V. D.; Trinkle, D. R.; Schmidt, N. W.; Davis, M. A.; DeVries, C.; Som, A.; Cronan, J. E., Jr.; Tew, G. N.; Wong, G. C. L. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 20595–20600. (33) Gabriel, G. J.; Maegerlein, J. A.; Nelson, C. F.; Dabowski, J. M.; Eren, T.; Nusslein, K.; Tew, G. N. Chem.;Eur. J. 2009, 15, 433–439. (34) Som, A.; Tew, G. N. J. Phys. Chem. B 2008, 112, 3495–3502. (35) Som, A.; Vemparala, S.; Ivanov, I.; Tew, G. N. Biopolymers 2008, 90, 83–93. (36) Scott, R. W.; DeGrado, W. F.; Tew, G. N. Curr. Opin. Biotechnol. 2008, 19, 620–627. (37) Lienkamp, K.; Madkour, A. E.; Musante, A.; Nelson, C. F.; Nusslein, K.; Tew, G. N. J. Am. Chem. Soc. 2008, 130, 9836–9843. 1348
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
Analytical Chemistry
ARTICLE
(38) Gabriel, G. J.; Tew, G. N. Org. Biomol. Chem. 2008, 6, 417–423. (39) Ilker, M. F.; Nusslein, K.; Tew, G. N.; Coughlin, E. B. J. Am. Chem. Soc. 2004, 126, 15870–15875. (40) Arnt, L.; Nusslein, K.; Tew, G. N. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 3860–3864. (41) Arnt, L.; Tew, G. N. J. Am. Chem. Soc. 2002, 124, 7664–7665. (42) Kuroda, K.; Caputo, G. A.; DeGrado, W. F. Chem.;Eur. J. 2009, 15, 1123–1133. (43) Chen, Z.; Shen, Y. R.; Somorjai, G. A. Annu. Rev. Phys. Chem. 2002, 53, 437–465. (44) Kim, J.; Somorjai, G. A. J. Am. Chem. Soc. 2003, 125, 3150– 3158. (45) Kim, G.; Gurau, M. C.; Lim, S.-M.; Cremer, P. S. J. Phys. Chem. B 2003, 107, 1403–1409. (46) Oh-e, M.; Hong, S.-C.; Shen, Y. R. Appl. Phys. Lett. 2002, 20, 784–786. (47) Ye, S.; Noda, H.; Nishida, T.; Morita, S.; Osawa, M. Langmuir 2004, 20, 357–365. (48) Zhuang, X.; Miranda, P. B.; Kim, D.; Shen, Y. R. Phys. Rev. B 1990, 172, 303–306. (49) Bain, C. D. J. Chem. Soc., Dalton Trans. 1995, 91, 1281–1296. (50) Eisenthal, K. B. Chem. Rev. 1996, 96, 1343–1360. (51) Wang, J.; Chen, C.; Buck, S. M.; Chen, Z. J. Phys. Chem. B 2001, 105, 12118–12125. (52) Wang, J.; Paszti, Z.; Even, M. A.; Chen, Z. J. Am. Chem. Soc. 2002, 124, 7016–7023. (53) Scatena, L. F.; Brown, M. G.; Richmond, G. L. Science 2001, 292, 908–912. (54) Gautam, K. S.; Schwab, A. D.; Dhinojwala, A.; Zhang, D.; Dougai, S. M.; Yeganeh, M. S. Phys. Rev. Lett. 2000, 85, 3854–3857. (55) Yang, C. S.-C.; Richter, L. J.; Stephenson, J. C.; Briggman, K. A. Langmuir 2002, 18, 7549–7556. (56) Bordenyuk, A. N.; Jayathilake, H.; Benderskii, A. V. J. Phys. Chem. B 2005, 109, 15941–15949. (57) Ma, G.; Liu, D. F.; Allen, H. C. Langmuir 2004, 2004, 11620– 11629. (58) Fitchett, B. A.; Conboy, J. C. J. Phys. Chem. B 2004, 108, 20255– 20262. (59) Ye, S.; Morita, S.; Li, G.; Noda, H.; Tanaka, M.; Uosaki, K.; Osawa, M. Macromolecules 2003, 36, 5694–5703. (60) Kweskin, S. J.; Komvopoulos, K.; Somorjai, G. A. Langmuir 2005, 21, 3647–3652. (61) Rivera-Rubero, S.; Baldelli, S. J. Phys. Chem. B 2004, 108, 15133–15140. (62) Even, M. A.; Lee, S.-h.; Wang, J.; Chen, Z. J. Phys. Chem. B 2006, 110, 26089–27097. (63) Baldelli, S. Acc. Chem. Res. 2008, 41, 421–431. (64) Wang, J.; Even, M. A.; Chen, X.; Schmaier, A. H.; Waite, J. H.; Chen, Z. J. Am. Chem. Soc. 2003, 125, 9914–9915. (65) Nguyen, K.; King, J. T.; Chen, Z. J. Phys. Chem. B 2010, 114, 8291–8300. (66) Nguyen, K. T.; Le Clair, S. V.; Ye, S.; Chen, Z. J. Phys. Chem. B 2009, 113, 12358–12363. (67) Ye, S.; Nguyen, K. T.; Le Clair, S. V.; Chen, Z. J. Struct. Biol. 2009, 168, 61–77. (68) Liu, J.; Conboy, J. C. J. Am. Chem. Soc. 2004, 126, 8376–8377. (69) Tong, Y.; Li, N.; Liu, H.; Ge, A.; Osawa, M.; Ye, S. Angew. Chem., Int. Ed. 2010, 49, 2319–2323. (70) Chen, C.; Clarke, M. L.; Wang, J.; Chen, Z. Phys. Chem. Chem. Phys. 2005, 7, 2357–2363. (71) Chen, C.; Wang, J.; Chen, Z. Langmuir 2004, 20, 10186–10193. (72) Lu, X.; Shephard, N.; Han, J.; Xue, G.; Chen, Z. Macromolecules 2008, 41, 8770–8777.
1349
dx.doi.org/10.1021/ac1025804 |Anal. Chem. 2011, 83, 1342–1349
