|Home | About | Journals | Submit | Contact Us | Français|
A family of peptoid dimers developed to mimic SP-B is presented, where two amphipathic, cationic helices are linked by an achiral octameric chain. SP-B is a vital therapeutic protein in lung surfactant replacement therapy, but its large-scale isolation or chemical synthesis is impractical. Enhanced biomimicry of SP-B’s disulfide-bonded structure has been previously attempted via disulfide-mediated dimerization of SP-B1-25 and other peptide mimics, which improved surface activity relative to the monomers. Herein, the effects of disulfide- or ‘click’-mediated (1,3-dipolar cycloaddition) dimerization, as well as linker chemistry, on the lipid-associated surfactant activity of a peptoid monomer are described. Results revealed that the ‘clicked’ peptoid dimer enhanced in vitro surface activity in a DPPC:POPG:PA lipid film relative to its disulfide-bonded and monomeric counterparts in both surface balance and pulsating bubble surfactometry studies. On the pulsating bubble surfactometer, the film containing the ‘clicked’ peptoid dimer outperformed all presented peptoid monomers and dimers, and two SP-B derived peptides, attaining an adsorbed surface tension of 22 mN m−1, and maximum and minimum cycling values of 42 mN m−1 and near-zero, respectively.
Lung surfactant (LS) is a complex lipid-protein mixture that lines the air-liquid (a/l) interface of vertebrate lungs and varies surface tension (γ) with the volume changes that occur during normal respiration. It reduces the work of breathing, maximizes lung compliance, optimizes the surface area available for gas exchange, and stabilizes the alveolar network against collapse1–4. Surfactant replacement therapy (SRT) is an effective, often animal-derived treatment for infant respiratory distress syndrome (IRDS)1, which occurs with significant incidence in infants born prior to 32 weeks of gestation, when the lungs are underdeveloped and do not secrete functional LS. Currently, there is no effective treatment for acute respiratory distress syndrome (ARDS), which affects adults after lung injury, infection, and in other instances2,3. Due to drawbacks associated with both animal- and synthetically derived SRT formulations4–7, a great, innovative opportunity remains to develop a fully functional, biomimetic LS of precisely known composition, that is cost-effective, biostable, and non-immunogenic, and that minimizes the difficulties of aggregation and protease-degradation associated with using peptides and proteins.
LS comprises by weight ~ 90% lipids (~ 75% of which is phosphatidylcholine) and ~ 10% combined protein, including surfactant proteins SP-A, SP-B, SP-C, and SP-D8–10. The three requirements for functional LS are: (1) rapid adsorption within one minute to the a/l interface, (2) attainment of near-zero γ at expiration, and (3) excellent re-spreadability of material at the interface during respiration, i.e., multiple continuous cycles of film expansion and compression. Although DPPC (a saturated, zwitterionic phospholipid) films reach near-zero γ upon compression, they fail to adsorb quickly or re-spread efficiently during cycling. The inclusion of additional, unsaturated or neutral lipids to the film improves re-spreading but raises the minimum γ reached11. It has been established that the two hydrophobic SPs, SP-B and SP-C, are critical for the γ-reducing and stabilizing behavior of natural LS, as they are involved in maintaining lipid organization, transport, and the formation of surfactant-associated reservoirs at the a/l interface12–14.
SP-B is a small, highly conserved 79mer with ten cationic residues, ~ 45% helicity, and four to five amphipathic helices that are presumably constrained by three intramolecular disulfide bonds and one intermolecular disulfide bond, the latter promoting homo-dimerization in its natural form15–18. No resolved structure is currently available. Extensive research into the in vitro and in vivo biophysical properties of SP-B has revealed that in vivo efficacy is improved when clinical formulations include SP-B, and debatably, in preference to the inclusion of SP-C5,19,20. Unfortunately, the unique and highly conserved structural characteristics of SP-B (and SP-C) prohibit practical large-scale isolation or chemical synthesis in a manner that preserves the structural integrity and stability, and hence function, of the protein when bound to lipids.
Logically, the design, synthesis, and characterization of simpler peptide mimics (KL4)21 or shorter SP-B fragments (SP-B1-25, SP-B53-78)22–26 have been explored. The latter terminal, surface-active segments are connected via two proximally close disulfide bonds in natural SP-B. This orientation is believed to facilitate transient insertion into the interfacial lipid film by electrostatic and hydrophobic interactions with the lipid head-groups and acyl chains, promoting stability and minimizing loss of material to the hypophase. A peptide mimic, dSP-B1-25, which links two monomers via a disulfide bond, has been extensively researched, and results have indicated that disulfide-mediated dimerization significantly increased surfactant activity for this peptide27–30. Additionally, an N-/C-terminus construct “Mini-B,” which links SP-B8-25 and SP-B53-78 with two disulfide bonds, also showed very promising surfactant activity relative to a mixture of the respective monomers, with performance as good as or better than the natural protein31,32.
The positive contribution to surfactant activity that is seemingly caused by disulfide-mediated dimerization of SP-B peptide mimics has led to the current investigative study. As previously outlined, poly-N-substituted glycines (peptoids) have recently been advanced as promising alternatives to peptide therapeutics because they are able to circumvent protease-susceptibility and irreversibly aggregative tendencies typical of hydrophobic proteins and peptides33,34. Briefly, peptoids are synthesized via a standard solid-phase protocol35 and contain nitrogen-substituted side chain variants of their peptide cousins, rendering the backbone achiral and hydrogen bond donor-deficient. However, through the inclusion of bulky, chiral side chains, sequence-specific peptoids are manipulated to adopt a preferred helical handedness, and can form extremely stable polyproline type I-like helices through steric and electronic repulsions36,37. Depending on the side chain properties, peptoids exhibit a pitch of ~ 6–6.7 Å and ~ 3 residues per turn38,39. Design and characterization of peptoids that mimic the generic structural attributes of relevant antimicrobial or SP-related sequences have confirmed their ability to mimic the structure, form, or function of the peptides40–45.
This begs the question as to whether a previously characterized peptoid that mimics the helicity, amphipathicity, and hydrophobicity of the N-terminus of SP-B will exhibit enhanced in vitro surface activity upon dimerization41–43. We report here the design, synthesis, and in vitro characterization of five dimerized peptoids. Four of these are dimerized using a disulfide linkage: one via the secondary amine at the N-terminus, and three others at the C-terminal end, flanked by an achiral octamer spacer of varying hydrophobicity. The fifth dimer replicates one of the dimers with a C-terminal linker, but the linkage is formed through ‘click-chemistry’ (1,3-dipolar cycloaddition) instead of a disulfide bond46,47. The ‘click-chemistry’ triazole-containing linkage has been shown to impart desired pharmacologic properties and activities to a variety of molecules48–50. To assess in vitro surface activity in a “Tanaka lipid” film (TL)51: 1,2-diacyl-sn-glycero-3-phosphocholine (DPPC): 1-palmitoyl-2-oleoyl-sn-glycero-3-[phospho-rac-(1-glycerol)] (POPG): palmitic acid (PA) 68:22:9 by weight, all peptoids were characterized using a Langmuir-Wilhelmy surface balance (LWSB) equipped with epifluorescent microscopic (FM) imaging and a pulsating bubble surfactometer (PBS). Monomeric variants of the dimers, with and without a linker, were also characterized for comparative purposes. For relevant techniques, the dimers and monomers were compared to established, previously characterized peptide mimics of SP-B (KL4 or SP-B1-25).
In this study, we find that disulfide-mediated dimerization led to little or no improvement in peptoid surface activity relative to the monomer. Enhanced surface activity depended on the side chain chemistry and presence of the linker. However, ‘click’-mediated dimerization significantly improved in vitro surface activity, seemingly by imparting conformational rigidity and polarity to the structure. The activity characteristics of the ‘clicked’ dimer match and even excel relative to those of current peptide mimics. This work represents the first instance of biophysical activity by a ‘clicked’ peptoid47, and may influence future peptide- and/or peptoid-based dimer designs.
The primary motivation for investigating peptide- or peptoid-based molecules as substitutes for SP-B in an LS formulation is the mounting evidence that SP-B plays a vital role in LS biophysical function. Unfortunately, SP-B is inaccessible in pure form on a large scale for use in a synthetic formulation. Replacing SP-B (or SP-C) with a fully functional molecule in a biomimetic formulation may result in a more bioavailable, cost-effective, and safer alternative than the currently used, animal-derived counterparts. To date, lone amphipathic peptide helices as SP-B mimics have failed to reproduce the behavior that results from the protein’s naturally more complex structure; although select peptides have shown very promising in vitro and in vivo LS activity, none have been introduced to the pharmaceutical market52,53. As a result, disulfide-mediated dimerization of peptide mimics has been extensively investigated in an attempt to enhance the biomimicry of SP-B’s sequence and structure. In particular, dSP-B1-25 and Mini-B significantly improved surfactant performance in both in vitro and in vivo tests when compared to their respective monomers.
Earlier generations of single helix peptoid-based SP-B mimics have displayed in vitro surface activity as good as or better than the SP-B1-25 and KL4 peptides41–43. The previously designed and characterized 17mer B1 (Figure 1) mimics the helicity, hydrophobicity, and cationic facial amphipathicity of both peptides, with particular attention paid to including the fewest number of different residues, the latter similar to KL4. Inclusion of bulky, chiral, and aromatic (Nspe, Phe-like) side chains in peptoids has increased in vitro surface activity relative to sequences with chiral, aliphatic (Nssb, Ile-like) side chains41,42. In the present work, the five dimerized peptoids presented in Figure 1 elaborated on the B1 sequence by exploring new avenues of SP-B biomimicry in non-natural oligomers.
Peptoid dB1 is a simple, N-terminally disulfide-bonded dimer of B1, which mimics one of the N-/C-terminus intramolecular disulfide bonds that naturally exists between SP-B1-25 and SP-B53-78. Peptidomimetic analogs dB2, dB3, and dB4 were also dimerized via a disulfide bond, but at their C-termini, and possess extended peptoid sequences, comprising an octameric bridging region that is primarily hydrophilic and achiral. The location of this disulfide bond mimics the intermolecular bond that homodimerizes two SP-B monomers, i.e., located positionally opposite the dimerized N-/C-terminus region. The octameric chain, which includes two four-residue segments that flank the disulfide bond, loosely mimics the SP-B23-54 peptide segment, which also flanks the intermolecular disulfide bond in the natural protein. In this SP-B segment, > 60% of the residues exhibit hydrophilic or mildly hydrophobic properties, with interspersed hydrophobic residues. In addition, large stretches of amino acids without a disulfide bond are present, implying a less structurally constrained region.
Therefore, the hydrophobicity of four (50%) of the side chains in the octameric linker region were increased for each molecule, comprising Nmeg, Nprp, or Npm in dB2, dB3, and dB4, respectively (Figure 1). Linker regions of varying side chain hydrophobicity may facilitate different degrees of interaction with the lipid film at the a/l interface, or may promote intra- or intermolecular association between two helices of the same or different units, respectively. Note that the extent of helicity of B1, previously reported41,42, is stabilized in peptoids through steric and electronic repulsions, and therefore, the addition of a C-terminal linker region that is primarily achiral would be expected to slightly reduce the helical signal, but have little to no effect on the secondary structure of the main peptoid helix (see Figure 1 SD).
To further explore the effect of dimer linkage on surfactant activity, azide- and alkyne-containing monomers of dB2 were conjugated via 1,3-dipolar cycloaddition, or ‘click-chemistry’ to form dB2c (Figure 1). The 1,4-disubstituted triazole ring has been shown to impart physiochemical stability and rigidity to different structures48. For comparative purposes, monomer variants of dB2 and dB3 were synthesized (mB2 and mB3), with an NCys→Nmeg substitution at the C-terminus (Figure 1) to prevent unwanted dimerization. Where applicable, peptoids were compared to the established peptide mimics SP-B1-25 and/or KL4.
LWSB experiments, performed at 25 and 37 C, yielded surface pressure (Π) - molecular area (Å2 molec−1, or A) isotherms for films spread at the a/l interface. Such isotherms characteristically exhibit a high (early) ‘liftoff’ area, the A at which Π first measurably increases from zero, a progression from the liquid-expanded (LE) to liquid-condensed (LC) phases, and finally, to a high collapse Π of ≥ 70 mN m−1 or near-zero γ. In addition, a pronounced, extended plateau often occurred at ~ 40–50 mN m−1. The first two characteristics correspond to the first two criteria for a functional LS film, i.e., a γ-reducing effect at the a/l interface (liftoff at a large film surface area), and reaching near-zero γ at compression (a high collapse pressure). It should be noted that although the liftoff feature is outside the physiologically relevant surface pressure regime for lung surfactant function (≥ 40 mN m−1), it is a useful indicator for determining the γ-reducing effect in the presence of the surface-active additive, and for estimating similarity to the behavior of SP-B29.
The third characteristic, an extended plateau in the isotherm, has a significance that is well-debated, but is believed to represent a phase transition corresponding to reversible squeeze-out of material (possibly of specific molecules) from the monolayer into associated surfactant reservoirs below the surface. This plateau may be related to re-spreadability and is a defining characteristic for the behavior of SP-B, SP-C, and mimics thereof, in the presence of lipids. Though the plateau size is a compression rate-dependent property, under the same compression rate, the relative plateau shapes and sizes of the lipid-peptoid films can be compared to that of the lipid-only film.
Lipid-peptoid isotherms, regardless of temperature, exhibited a high collapse Π ≥ 70 mN m−1 (Figure 2) which is mimetic of SP-B in this lipid mixture29, and of the natural LS film. For average liftoff data and 2D isotherm features ± standard deviation of the mean (σ), see Table I. At both temperatures, the Tanaka lipid (TL) film exhibited a later liftoff (lower molecular area) and smaller plateau than any of the lipid-peptoid films (93 Å2 molec−1 at 25 °C, 101 Å2 molec−1 at 37 °C, see Table I). At 25 °C (Figure 2, A & B), TL + B1 and + dB1 (the linker-free dimer of B1) had nearly overlapping isotherms, as did most of the other monomer/dimer pairs with the exception of the Npropyl-containing pair (TL + mB3/dB3) and the purely Nmeg-containing ‘clicked’ dimer, relative to its monomer (TL + mB2/dB2c).
Although TL + mB2 had an earlier liftoff than the film containing its disulfide-linked dimer dB2, the gradual slope rendered the difference unnoticeable in this regime, while the film containing the clicked variant dB2c had a significantly later liftoff. TL + dB4 (Npm-containing linker), while not shown, exhibited an isotherm that directly overlaid that of TL + B1. TL + dB3 (Npropyl) had the earliest liftoff, even above the film containing its monomer (mB3), while the dB2c-containing film had the latest (Table I). At a compression rate of 30 mm min−1, the plateau sizes were similar for all films except for the Npropyl-containing pair (TL + mB3/dB3), for which they were the longest in A (Table I). Notably, the isotherm shapes were different for this pair, where the lipid-dimer isotherm had a steeper slope and is therefore considered less compressible than the lipid-monomer film.
In a more fluid lipid film at a more physiologically relevant temperature, 37 °C (Figure 2, C & D), all isotherms exhibited earlier liftoffs and more pronounced plateaus than at 25 °C. Interestingly, with the exception of TL + dB2c, all of the lipid-dimer isotherms had earlier liftoffs than their respective lipid-monomer isotherms (Table I). Again, the film containing the dimer with Npropyl chains, TL + dB3, possessed the earliest liftoff and longest plateau, where an early liftoff and pronounced plateau are considered desirable surface activity characteristics that mimic SP-B in this lipid mixture29. The film with the clicked dimer (TL + dB2c) had a nearly overlapping isotherm with the lipid-monomer variant film, and as at 25 °C, TL + dB4 (Npm-containing, not shown) exactly overlaid TL + B1.
According to the listed criteria for functional LS, any one of these molecules could be a considered a good, surface-active mimic of SP-B. However, subtle but significant differences in surface activity characteristics lend more insight into the mechanics of these lipid-peptoid films. The addition of a peptoid + linker (mB2) did little to improve activity relative to B1. Dimerization and linker hydrophobicity had a much more significant effect on surface activity, primarily at 37 °C. N-terminal dimerization (dB1) with no linker did improve surface activity relative to the monomer, and C-terminal dimerization with the addition of a hydrophilic linker containing 50% aliphatic properties (dB3) seemed to have the optimal effect. To further evaluate the surface activity of these films, the FM film morphology was examined.
FM images were recorded at 25 and 37 °C on a Langmuir trough with 0.50 mol% Texas Red DHPE (TR-DHPE) spiked into the solution containing lipids or lipids + peptoid (Figure 3). TR-DHPE is a bulky, fluorescently headgroup-labeled lipid that is preferentially excluded from the more ordered regions of the monolayer upon film compression. Therefore, dark regions of the monolayer represent LC domains, while lighter regions correspond to the more fluid LE phase. Bright spots present in images may represent sub-monolayer vesicles or protrusions above the monolayer that are still associated with the film, and are termed here bright protrusions. It is not known whether bright protrusions contain a higher proportion of dye, or are folds with a multi-component composition; they are simply considered to be pockets of material removed from the monolayer.
FM images are presented at Π that correspond to before and after the plateau region at both temperatures: 41 and 51 mN m−1 at 25 °C (Figure 3, left image set), and 42 and 54 mN m−1 at 37 °C (Figure 3, right image set). For these films, LC domains formed at 3–7 mN m−1 at 25 °C and 25–30 mN m−1 at 37 °C; bright protrusions, if formed, occurred at 42–48 mN m−1 at both 25 °C and 37 °C. At 41 mN m−1 at 25 °C, the TL film contained LC domains surrounded by the brighter fluid LE phase. This morphology was mostly consistent for all of the lipid-peptoid films, where TL + mB2/dB2/dB2c had slightly smaller domains than TL, while TL + mB3/dB3 had slightly larger domains, and TL + B1 and TL + dB4 had very similarly sized domains to TL. Upon closer inspection, small bright protrusions were observed in TL + dB3 and TL + dB2c.
Interestingly, at 25 °C, the bright protrusions in TL + dB2c were seemingly located in or on the center of the LC domains. At 51 mN m−1, the TL film contained LC domains of roughly the same density as at 41 mN m−1, but with the additional inclusion of small bright protrusions. However, depending on the run, these bright protrusions were sometimes very sparse or seemingly not present. All of the lipid-peptoid films contained LC domains that were roughly the same size and density as those observed at 41 mN m−1. The major differences in film morphology were thus in bright protrusion formation and patterning, which for TL + dB1, TL + dB2, and TL + dB4, were very sparse or absent. Lipid-monomer films (TL + B1, TL + mB2, and TL + mB3) all exhibited bright protrusions, with most prevalence in the TL + mB3 film. TL + dB3 and TL + dB2c were the only lipid-dimer films that exhibited a significant amount of bright protrusions at this lower temperature. Though TL + dB2c bright protrusions were more sparse, their location in the center of the LC domains was more distinctly observed at 51 mN m−1.
At 37 °C, the domain boundaries in the FM images became less clear as they were more difficult to capture due to increased x-y translational movement (in-plane) of the more fluid film (Figure 3). At 42 mN m−1, the same trend in LC domain size was observed as at 25 °C. One notable difference was that for TL + dB2c, elongated rod-like shapes formed in addition to circular LC domains. These rod-like shapes are sometimes seen in the imaging of giant unilamellar vesicles (GUV’s), and have been established as evidence of a more fluidized film54. At 54 mN m−1, LC domains nearly disappeared in all lipid-monomer films, and for TL + dB4. Bright protrusions were present in all films, where the patterning was, in some cases, quite distinct. TL + B1 displayed very small, bright protrusions, whereas TL + dB1 had duller, larger bright protrusions. While TL + dB2 had sparse, bright protrusions, its lipid-monomer variant film (TL + mB2) displayed many more, larger, duller bright protrusions. Interestingly, TL + dB2c bright protrusions were either bright and sparse, or were duller and seemed to lie directly on the rod-like structures. TL + mB3/dB3 both had a tremendous amount of small, bright protrusions.
By correlating isotherm features with FM images, a better in vitro surface activity assessment for this technique is provided. Relating structure to function, N-terminal, simple dimerization did not significantly alter the phase morphology, and C-terminal dimerization resulted in morphology changes dependent on linker hydrophobicity. From the isotherms, the most surface-active films would be TL + mB3 or TL + dB3, and FM images revealed a possible cause for this increase in activity. An easily compressed, easily re-spread monolayer would likely display homogeneously sized, small LC domains that coexist with small bright protrusions at high Π. This promotes elasticity and facile reorganization of the monolayer into highly compressed states. Small bright protrusions are more likely to easily re-spread than larger sub-or super-monolayer structures, and for these two films, this was exactly observed. Note that the presence of a plateau region does not necessitate the presence of bright protrusions. A lack of bright protrusions may point to inferior surface activity or an inefficient mechanism of folding to accommodate higher surface pressures.
The frequency and density of bright protrusions coexisting with LC domains indicates that material was evenly removed, but still associated with, the interfacial monolayer, and may be re-spread upon expansion. Although not shown, all LC domains and bright protrusions disappeared with expansion of the compressed film. Surprisingly, the type of dimerization did seem to affect the phase morphology, where the most distinct patterning was observed with TL + dB2c. The unique shape of the LC domains and localization of bright protrusions indicates that the mechanism of material removal and reorganization upon film compression may be somewhat different in this case.
The adsorption kinetics of the lipid-peptoid aqueous suspensions were studied using the PBS at 37 °C in static mode, which monitors γ over time (Figure 4, A). A clinically administered LS formulation, Infasurf®, adsorbed rapidly to the interface in less than one minute, and attained a low equilibrium γ (γeq) of ~ 23 mN m−1 or less on the PBS55. Rapid adsorption to the a/l interface is one of the key criteria in LS biophysical functioning. The adsorption curves for lipid-peptoid films are presented in Figure 4, and the average γ data ± σ at selected time intervals, in comparison to SP-B1-25 and KL4, is available in the Supplemental Data (Table II SD). The TL film adsorbed quite slowly to the interface, requiring nearly 20 minutes to reach a high γeq of ~ 54 mN m−1. All lipid-peptoid films improved upon these adsorption characteristics, but to significantly different extents. TL + B1/dB1 reached relatively similar γeq values (39 and 41 mN m−1, respectively), but TL + dB1 adsorbed to this value much more quickly than TL + B1. TL + mB2 promoted more rapid adsorption to a significantly lower γeq than TL + dB2 (40 and 46 mN m−1, respectively). TL + dB2c reached a dramatically lower γeq of 22 mN m−1, which nearly matched that of KL4 and Infasurf®55, albeit at a slower adsorption rate. TL + mB3/dB3 adsorbed at more rapid rates and reached similar γeq values (37 and 38 mN m−1, respectively), which were improved over any of the other disulfide-bonded dimers or their monomeric counterparts. TL + dB4, although not shown, adsorbed more rapidly than TL + B1, but reached a significantly higher γeq of 46 mN m−1 (Table II SD). Even with slower adsorption – reaching γeq in five minutes – the TL + dB2c film demonstrated superior adsorption characteristics relative to the other lipid-peptoid films.
Perhaps the most physiologically relevant in vitro test for any LS formulation is the endurance of the film at the a/l interface during rapid changes in volume or surface area. Running the PBS in dynamic mode allows for a simplified evaluation of film behavior under such conditions. The γ-surface area (SA) data loops for each lipid-peptoid film, shown in Figure 4 B–C, represent one pulsation cycle at 20 cycles per minute (cpm) after five minutes, with expansion in a clockwise loop direction. The average γmax/min data ± σ at selected time intervals, in comparison to SP-B1-25 and KL4, is presented in Table II. Low γ data is absent, i.e., difficult to obtain, in some loops due to the inability of the image analysis system to trace bubble shape in this regime. At low γ, the lipid film is in a highly compressed state, and the bubble shape often deforms significantly from that of an ellipse or sphere, where the ellipsoidal Laplace equation is normally used to calculate γ and SA based on a trace of bubble shape and voltage data from the instrument. Regardless, it was clear from data analysis and visual, real-time inspection that γ reached near-zero in these films. Additionally, bubble size was not uniform for every experiment, but differences in absolute surface area (positioning on the x-axis, Figure 4) had no appreciable effect on γ as long as data could be accurately traced.
Maintaining a reduced γ throughout respiration, and importantly, near-zero γ upon film compression, are key features of LS that indicate film stability and re-spreadability. The PBS loop for Infasurf®55 is known to have a maximum γ (γmax) of ~ 35 mN m−1 and a minimum γ (γmin) near zero; near-zero γmin must be attained immediately upon cycling and should remain for an extended period with minimal compression. Although the significance of bubble hysteresis has not been established, a large amount of hysteresis often correlates with a low amount of compression to reach near-zero, which is a very desirable characteristic in LS formulations. The TL film exhibited a high γmax ~ 63 mN m−1 and γmin ~ 10 mN m−1, with ~ 44% compression (Figure 4, inset table, and Table II).
The addition of any one of the peptoid SP mimics to the lipid film resulted in a γmin near zero immediately at cycling commencement, a decrease in γmax, and a reduction in percent compression. The near-zero γmin is expected given the presence of this feature in previous characterization of TL + B1 films41–43. Although TL + B1/dB1 exhibited similarly shaped loops, TL + dB1 reached a higher γmax (57 mN m−1 relative to 52 mN m−1) and had an increased percent compression (39% relative to 31%). This trend was followed for every monomer/disulfide dimer pair, where the γmax was higher and percent compression increased (Table II). TL + dB4, although not shown, displayed activity and a loop shape very similar to TL + dB2. Of the disulfide-bonded dimers, TL + dB3 had the lowest γmax (53 mN m−1) and percent compression (36%), and of the monomers, TL + mB3 was the most surface-active, with a slightly lower γmax and about the same percent compression as TL + B1. However, the TL + dB2c film surprisingly demonstrated superior in vitro surface activity relative to all other dimers, monomers, and the KL4 peptide, with a γmax of 42 mN m−1 and 21% SA compression to reach 20 mN m−1.
These results indicate that N-terminal disulfide-based dimerization of peptoids without a linker renders little benefit to dynamic surface activity characteristics in a mixed lipid film. This is in direct contrast to the single disulfide-bonded dSP-B1-25, which was reported to be significantly more surface-active than the monomer. However, inclusion of a ≥ 50% hydrophilic linker region at the C-terminus in peptoids did yield some benefit, which depended on side chain chemistry and not dimerization. The molecules that possessed 50% aliphatic residues, mB3 and dB3, performed the best in their respective classes (monomer and dimer). Interestingly, with the exception of dB3, the addition of a disulfide-based dimer to the lipid film resulted in inferior surface activity relative to all of the monomers, including the original monomer B1. However, a dramatic improvement was witnessed when a peptoid dimerized via ‘click-chemistry’ was added to the lipid film, resulting in dynamic film surface activity characteristics that surpass all other peptoids and both single helix peptide mimics KL4 and SP-B1-25.
In this study, we explored the effects of dimerization and linker hydrophobicity on in vitro surface activity of peptoid mimics of SP-B in a mixed lipid film. The presence of multiple disulfide bonds in natural SP-B facilitates the in vitro and in vivo surface activity that enables the lung’s biophysical function. Three intramolecular disulfide bonds, two connecting the N-/C-terminus segments, and an additional one further along in the sequence, constrain the secondary structure and flexibility of the protein when bound to lipids. For SP-B homodimers in a surfactant film, the intermolecular disulfide bond is postulated to act as a hinge, effectively anchoring, organizing, and inserting squeezed-out lipid assemblies to and from the monolayer at the alveolar interface. This activity is believed to be vital for efficient re-spreading of material throughout the alveolar area cycling that occurs during breathing.
Whether these disulfide bonds are absolutely required for protein function has been extensively studied, but their exact contributions to molecular structure and function in vivo remain unknown. Although chemically synthesized SP-B1-78 has exhibited good in vitro surface activity, it was still inferior when compared to the naturally isolated protein29,56. Biophysical functioning of natural SP-B was inhibited if the protein was completely reduced57, or if homo-dimerization was prohibited transgenically or through the position-specific substitution of Cys48 → Ser in the sequence58,59. Although totally reduced forms of natural SP-B demonstrated surface activity, the protein behaved in a manner that is congruent with increased structural flexibility57.
Further evidence that supports the necessity of dimerization is the propensity of SP-B and other saposin-like proteins to form a buried hydrophobic cavity60,61. This cavity has been postulated to partake in lipid binding events that aid in the lipid transport and reorganization at the interface during respiration. This buried domain may form during oligomerization or through hydrophobic and hydrophilic associations between SP-B monomers, but disulfide bonds likely ensure the presence of this cavity through conformational restrictions.
The current study elucidates possible differences between the mechanisms of surface activity in peptide and peptoid mimics of SP-B. Whereas disulfide-mediated dimerization of SP-B1-25 near the N-terminus resulted in improved surface activity in vitro and in vivo when compared to the monomer27, peptoids were significantly less affected. Interestingly, on the trough at 37, but not 25 °C, disulfide-bonded dimers exhibited earlier (higher) liftoff areas than the monomers. In the more fluid film at an elevated temperature above the Tc of the lipid mixture, more peptoid dimer may have been able to insert than at lower temperature, thus decreasing the available space per molecule at the interface. Because the dimers are roughly twice as large as the monomers, this may have translated to an earlier liftoff at a higher area per molecule. However, a seemingly opposite trend in surface activity was observed on the PBS at 37 °C, where the monomers tended to have faster adsorption rates and lower γeq, in addition to a lower γmax, than the disulfide-based dimers. Because the LWSB is a quasi-equilibrium, slow-cycling technique, isotherm features cannot always be directly correlated to adsorption and cycling features on the PBS. In the rapidly cycled dynamic film on the PBS, it appears that disulfide-mediated dimerization may have actually hindered surfactant activity in peptoid mimics. Thus, for disulfide-based peptoid mimics, it appears that early liftoff areas do not always correlate directly with an increase in surface activity observed when using more physiologically relevant techniques.
As SP-BCys48 → Ser monomers have been shown to associate as dimers non-covalently59, which contributed favorably to their activity, mimicking the natural sequence, there exists some evidence that peptoid oligomerization of “monomer” (helix) units can occur in salt solution through associations of the aromatic side chains (Yoriel Marcano, Annelise E. Barron, unpublished work). This phenomenon could explain the lack of improvement observed with disulfide-mediated dimerization of B1. If close proximity of the two peptoid helices to each other is the most important factor in contributing to surface activity, and extensive peptoid oligomerization is already occurring, disulfide-mediated dimerization could be unnecessary, or, as observed on the PBS, inhibiting. Translating to the biophysical function of lung surfactant, an underlying mechanism of aromatic peptoid-peptoid intermolecular associations at the a/l interface may promote sufficient film organization/folding and lipid transport during surface area expansion and compression, rendering a simple disulfide bond of little benefit.
Although mostly excluded from peptide mimic sequences, the > 60 % hydrophilic residues in SP-B23-54 may also play a critical role in lipid insertion and organization at the a/l interface62. By introducing an achiral, octameric peptoid chain with ≥ 50% hydrophilic residues at the C-terminus of B1 in mB2 and mB3, we attempted to mimic this portion of the protein. Results indicated that the in vitro surface activities of each of the peptoids in the lipid film depended more on the side chain chemistry of the linker than on C-terminal, disulfide-mediated dimerization.
Differences in the phase morphology of lipid-peptoid films were attributed mostly, to linker hydrophobicity and not disulfide-mediated dimerization. At lower temperature and higher Π, the lack of surface activity in the disulfide-bonded dimers, except for dB3, was clearly observed, where inclusion into the lipid film did not result in bright protrusion formation, which was not the case for inclusion of any monomer. The absence of bright protrusions despite the presence of a plateau region in these films could result from a different mechanism of material removal that is less efficient for reaching near-zero surface tensions and subsequent re-spreading. This may translate physiologically to increased re-spreading upon film expansion (inhalation) and increased formation of sublayer structures at compression (expiration). Only the presence of mB3 and dB3 resulted in extensive bright protrusion formation, which corroborated the increased activity observed in the isotherms (early liftoff and an extended plateau) in this Π regime.
However, at higher temperature and higher Π, the disulfide-based dimers became more surface-active, as indicated by extensive bright protrusion patterning in all cases. Though again, only the inclusion of N-propyl chains in mB3 and dB3 resulted in the desired formation of homogeneous and small bright protrusions with high density and frequency. This result further supports the idea that N-propyl chains are able to facilitate hydrophobic interactions with other peptoid residues, or partially insert into and associate with the lipid acyl chains at the a/l interface. A purely hydrophilic linker (Nmeg) would prefer to associate with the lipid headgroups or aqueous buffered subphase, while inclusion of benzyl chains (Npm) could contribute a bulkiness to the linker that could exclude it from any activity-enhancing interactions with the tightly packed lipid film. Interestingly, LC domains largely disappeared at higher Π when any monomer was added, indicating that single aromatic peptoid helices do not support or stabilize LC domains at higher Π, as the dimers do.
The advantages of including 50% aliphatic residues in the linker region are further supported by the PBS adsorption and dynamic cycling data. While inclusion of dB1, mB2, dB2, and dB4 led to little to no improvement in adsorptive properties relative to B1, both mB3 and dB3 enabled a more rapid adsorption of the lipid film, and created surface films that reached a lower γeq than was seen with the other monomers and disulfide-based dimers. It has been previously shown in peptoid mimics of SP-B that inclusion of aliphatic residues, and in general, an increased proportion of hydrophobicity, facilitated significantly more rapid adsorption to the a/l interface42,44. Of the disulfide-based dimers, dB3 resulted in the lowest γmax and a slightly lower % compression, where the N-propyl chains may once again promote better association with the lipid acyl chains of the dynamic film. Overall, the properties of the achiral portion of the peptoid linker did affect the biophysical activity of the lipid-peptoid film, with results pointing to changes in the ability of peptoid to aid in the folding and exclusion of lipids at higher surface pressures, which if inferior, could potentially lead to decreased re-spreading and higher maximum surface tensions during respiration.
Ultimately, the in vitro dynamic surface activity performance of dB2c in a lipid film far surpassed that of any of the disulfide-based dimers or monomers. The triazole moiety is known to be rigid, highly stable, and polar, where nitrogen atoms two and three function as weak hydrogen bond acceptors48. The triazole ring is also conformationally more restrictive than a simple disulfide. In previous work, incorporating a triazole amino acid with two flanking methylene groups into a peptoid oligomer caused a hairpin turn, effectively altering the conformation of the structure63. We hypothesize that the dramatic improvement in activity could be attributed to the conformational restrictions that this linkage imposes on the terminal peptoid helices and the linking unit. In addition, the late (low) liftoff areas of the dB2c-containing film, at both temperatures on the LWSB, support our previous interpretations about dimer exclusion from the lipid film. This could indicate conformational restriction, where the peptoid was physically unable to occupy sufficient area at the interface to affect liftoff. However, it is also possible that the inclusion of four flanking, freely rotatable methylene groups in both the disulfide and triazole linkages may negate the restrictive effect of the triazole group.
Alternatively, the electronic/physical properties of the triazole (an aromatic, electron-donating group) may be another explanation for the observed increase in surface activity. One could envision associations with the nearby C-terminal amide protons or perhaps one of the NLys protons through intra- or intermolecular hydrogen bonding. In addition, the electron-rich triazole ring could be involved in a cation-π interaction with an NLys ammonium cation64. Restricted linkage, potential hydrogen bonds, or cation-π interactions could all contribute to lowering the overall entropy of the molecule, and thereby facilitate a helical conformation that more favorably interacts with the anionic lipid headgroups of the lipid-peptoid film, thus enhancing biophysical activity.
From isotherm features alone, the activity of dB2c seemed very typical, but FM imaging of the lipid film revealed a significant and unique effect on phase morphology. The difference in LC domain shape at higher temperature, particularly with the formation of well-distributed, elongated rods, and the unusual ‘attachment’ of bright protrusions to the LC domains at both temperatures, indicated a fluidizing effect and a mechanism of material removal from the interface that was distinct and unique from the other lipid-peptoid films. It is also possible that the linkage acts as a polar anchor to the lipid headgroups, unable to insert into the lipid acyl chains, but nonetheless affecting phase morphology. Pursuing a more physiologically mimetic surfactometry technique (PBS) with this molecule revealed excellent adsorptive and dynamic film properties, relative to all other peptoids and established peptide mimics studied. The slightly slower adsorption of the lipid film in the presence of dB2c relative to KL4 can be attributed to the hydrophilic nature of this molecule; perhaps inclusion of 50% N-propyl chains would rectify this issue.
Therefore, single, amphipathic, cationic and aromatic peptoid helices were significantly more affected by a dimerization connectivity that either conformationally or electronically/physically changed the overall structure. From our work, it appears that disulfide-mediated dimerization had little effect on helicity. Single disulfides are known to be fairly flexible, with free rotation of the terminal peptoid helices about the bond. Evidence that conformationally restrictive linkages (like that possible for the triazole-containing linker) improve surface activity is present in the literature, where Raman/ATR-IR of the disulfide bond in dSP-B1-25 vs. dSP-B8-25 (dCys8) demonstrated that inclusion of the flexible SP-B1-9, which flanked the disulfide bond, resulted in a conformational restriction to a higher energy λ(C-SS-C) dihedral angle with increased strain30. Although the bond in dSP-B8-25 resided predominantly in a lower energy λ(C-SS-C) conformation, it was determined that inclusion with lipids caused both molecules to favor the conformation adopted by dSP-B1-2530. Furthermore, NMR studies of reduced and oxidized “Mini-B” have demonstrated that the two disulfide bonds promoted close, almost parallel association of the helices, and a very amphipathic structure32. The reduced form of Mini-B possessed a less defined C-terminal helical structure as well.
Although it is not well understood at this time which specific molecular properties of SPs control the stability and activity of a dynamic interfacial surfactant film experiencing cycling of alveolar surface area, it is clear that the same structure-based additions that increased surface activity in peptides may not correspond to increased surface activity in peptoids. The molecular mechanism and structural differences between peptoids and peptides are not sufficiently well-defined to speculate on reasons for disparities in surface activity upon disulfide-mediated dimerization. Two important features missing in peptoids are backbone hydrogen bonding, which is known to stabilize peptide helices, and an overwhelming preference for the trans form in cis/trans isomerization of peptide bonds. The cis conformation in amide bonds linking peptoid residues is known to be more highly populated than in peptides65. These two factors could contribute to the increased activity observed after disulfide-mediated dimerization in peptides, where the resulting folded dimeric structure may be more conformationally stable and better able to interact with lipid head groups than its peptoid counterpart. Regardless, we are intrigued by and will continue to investigate the improvements in biologically relevant activity that ‘click’-mediated dimerization seemed to impose on peptoids for lung surfactant applications, in both in vitro and in vivo environments.
Working toward our broader vision to create a fully functional biomimetic lung surfactant formulation comprising peptoid mimics of SP-B and SP-C, we have created a family of dimerized peptoids that mimic the in vitro surface activity of SP-B by mimicking its specific structural attributes. By exploring dimerization and linker hydrophobicity in these molecules, we have established that a) dimerization via ‘click-chemistry’ is far more beneficial to surface activity than disulfide-mediated dimerization in peptoids, b) the potential of peptoids to oligomerize in a hypophase-like environment may explain the lack of activity improvement in disulfide-bonded mimics, and c) including an achiral, hydrophilic linker with 50% aliphatic residues does render some surface activity improvement. We have provided support for the theory that conformational restriction of secondary and tertiary structure in SP-B may be vital to its biophysical function in the lung. In a broader context, we have demonstrated that ‘click-chemistry’ may be a viable option for developing natural and non-natural peptide mimics of SPs as well as other biological molecules that primarily operate in a lipid film. The triazole-containing moiety is more rigid and stable and contains more advantageous electronic/physical properties than a disulfide bond, which may render a degree of improved feasibility to drug development in this field. Additionally, this work has further cemented the idea that a completely synthetic, non-natural molecule can adopt structures and exhibit functional physical activity similar to its natural counterparts, which escalates the potential for these molecules to be eventually utilized in a therapeutic setting.
Peptide and peptoid synthesis reagents and supplies were purchased from Applied Biosystems (ABI) (Foster City, CA) and Sigma-Aldrich (Milwaukee, WI). Fmoc-protected amino acids and resins were purchased from EMD Biosciences (NovaBiochem, San Diego, CA). Primary amines (highest % purity and enantiomeric excess available), di-tert-butyl dicarbonate (Boc), and triphenylmethanol (Trt) were purchased from Sigma-Aldrich. Acetonitrile (ACN), chloroform, methanol, and trifluoroacetic acid (TFA), HPLC grade or better, were purchased from Fisher Scientific (Pittsburgh, PA). All salts were purchased from Fisher Scientific. DPPC and POPG were purchased from Avanti Polar Lipids (Alabaster, AL) and used as received. Palmitic acid (PA) was purchased from Aldrich. Texas Red®, 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine, triethylammonium salt (TR-DHPE) was purchased from Molecular Probes (Eugene, OR). All chemicals were used without further purification. Water was Milli-Q 18.2 mΩ·cm quality.
The two peptides, modified SP-B1-25 (Cys8,11 → Ala) and KL4, were synthesized by standard SPPS66 Fmoc chemistry on a 0.25 mmol scale using preloaded Wang resin and an ABI 433A automated peptide synthesizer. Peptoids were synthesized by the submonomer method35 using Rink amide resin, on a 0.25 mmol scale, and an ABI 433A, with Boc protection of N-(4-aminobutyl)glycine (NLys), and Trt protection of N-(mercaptoethyl)glycine (NCys). All molecules were cleaved from their respective resins by agitation in 90–95% TFA/water (v/v), along with the appropriate scavengers, for 5–10 minutes (peptoids) to one hour (peptides). Crude product for purification was obtained by immediate resin filtration of the mixture, dilution with ACN/water, repeated lyophilization, and re-dissolution in ACN/water. All molecules were purified on a Waters (Waters Corp., Milford, MA) or Varian Prostar (Varian, Inc., Palo Alto, CA) reverse-phase high performance liquid chromatography (RP-HPLC) system with a Grace Vydac (Deerfield, IL), Peeke Scientific (Redwood City, CA), Phenomenex (Torrance, CA), or Varian Dynamax C4 or C18 column, using a linear gradient of % solvent B in % solvent A over a selected time period (solvent A is 0.1% TFA in water [v/v] and solvent B is 0.1% TFA in ACN [v/v]), using standard purification techniques.
Disulfide-dimerized peptoids were synthesized and purified first as monomers. Disulfide-mediated dimerization occurred through 24hr air oxidation in an agitated ACN/water solution, at room temperature, and was monitored by analytical RP-HPLC. The conjugation of peptoid monomers via ‘click-chemistry’ was performed by first synthesizing and purifying the otherwise symmetrical azide and alkyne peptoid monomers (see SD), and then conducting a microwave-assisted ‘click’ reaction (100 °C, 1 h) using a Biotage Initiator microwave synthesizer system (Uppsala, Sweden), where known concentrations of both monomers were stirred in t-BuOH/H2O (1:1) containing 5 mol% CuSO4·5H2O and 15 mol% L-(+)-sodium ascorbate (see SD)49. The reaction mixture was dialyzed using a Slide-A-Lyzer dialysis cassette (Pierce, Rockford, IL) with a cutoff size of 3.5 kDa, and then RP-HPLC purified. All molecules were obtained in final pure form as a repeatedly lyophilized powder. Final purities were confirmed to be > 97% by analytical RP-HPLC and molecular weights were obtained by either electrospray ionization mass spectrometry (ESI/MS) or matrix-assisted laser desorption ionization time of flight mass spectrometry (MALDI-TOF/MS) (Figure 1, inset table).
The lipids DPPC, POPG, and PA were individually dissolved in a chloroform/methanol solution (3/1 [v/v]) to a known concentration (~ 2 or 4 mg mL−1). Single-lipid solutions were then combined by volume at the ratio of DPPC:POPG:PA, 68:22:9 [w:w:w] and to a known total lipid concentration (~ 2 mg mL−1). This well-characterized and well-known lipid formulation is considered an adequate mimic of the non-protein (lipid) fraction of LS51. The peptides and peptoids were individually dissolved in methanol from a lyophilized powder to a known concentration (1–2 mg mL−1). For the surfactometry studies, the peptides and peptoids were ‘spiked’ to the lipid mixture at 10 wt% relative to the total lipid content (~ 9–10 absolute wt%, see Table I SD), and to a final concentration of ~ 1 mg lipid mL−1. Based on mole % (Table I SD) or calculated number of molecules, there is a roughly equal number of simple monomer unit helices in the B1 and simple dimer, dB1, surfactant samples. Note that previously published work with B1 at a slightly lower mole percentage (2.16 mol%) resulted in improved PBS static-mode adsorptive properties41–43.
Π-A isotherms were obtained using a custom-built LWSB, which has been previously described in detail41. The trough was filled with 300 mL of aqueous buffer (150 mM NaCl, 10 mM HEPES, 5 mM CaCl2, pH 6.9) and heated to 25 or 37 °C. A Wilhelmy plate (Reigler & Kirstein GMBH, Berlin, Germany) was used to monitor surface pressure and was calibrated in buffer before each run. Each sample was spread at the a/l interface from a chloroform/methanol solution using a glass syringe and allowed to equilibrate for 5–10 minutes. The barriers were then compressed, expanded, and compressed again at a rate of 30 mm min−1. Isotherm measurements were repeated a total of six times for repeatability and to eliminate effects due to any possibility of leakage. Represented isotherms are first compressions displayed in Figure 2, and averages of important phase transition markers ± σ are presented in Table I.
To record FM images, a Nikon MM40 compact microscope stand with a 100W mercury lamp (Tokyo, Japan) was used in conjunction with the Langmuir trough. Epifluorescence was detected by a Dage-MTI three-chip color camera (Dage-MTI, Michigan City, IN) in conjunction with a generation II intensifier (Fryer, Huntley, IL). Samples were spiked with 0.50 mol% TR-DHPE, a fluorescently headgroup-labeled lipid, for detection. Isotherm features remained unchanged after TR-DHPE addition, and presumably, film morphology was largely unchanged by the presence of TR-DHPE at these concentrations. FM images were acquired directly from the compressed film on the a/l interface. Experiments were conducted exactly as the LWSB studies of un-spiked films, with the exception that barrier speed was reduced to 5 mm min−1 and experiments were repeated three times for repeatability. Average domain sizes were calculated using ImageJ software (The National Institutes of Health, Bethesda, Maryland).
A commercial PBS instrument (General Transco, Largo, FL), modified with a direct, real-time imaging system, which has been previously described and validated in detail55, was utilized to obtain both static-mode and dynamic-mode data. Samples were dried from chloroform/methanol 3/1 [v/v] in Eppendorf tubes using a DNA 120 speedvac (Thermo Electron, Holbrook, NY), forming a pellet. The pellet was suspended in buffer (150 mM NaCl, 10 mM HEPES, 5 mM CaCl2, pH 6.9) to 1.0 mg lipid mL−1, with a final volume of ~ 70 μL. The samples were then mixed with a pipette 20 times, sonicated with a Fisher Model 60 probe sonicator for two 15 second spurts, and then mixed again 20 times to form a dispersed suspension. Samples were then loaded into a small plastic sample chamber (General Transco) using a modified leak-free methodology55,67. The sample chamber was then placed in the instrument, surrounded by a water bath held at 37 °C. A bubble with a radius of 0.4 mm was then formed, and surface area was monitored throughout the experiment (bubble size gradually increased in both data collection modes, but had a negligible effect on γ).
Static-mode adsorption data were collected for 20 minutes, where the suspension was allowed to adsorb to the bubble surface over time. Adsorption data were smooth fit to a curve in the Kaleidagraph program by applying a Stineman function to the data, where the output of this function then had a geometric weight applied to the current point and ± 10% of the data range to arrive at the smoothed curve (Figure 4). Average γ ± σ at selected time intervals are listed in Table II SD. Dynamic-mode data were then subsequently obtained for each sample at the adult respiratory cycle frequency of 20 cpm for 10 minutes, with a 50% reduction in surface area per pulsation cycle. PBS experiments were repeated six times for each sample to ensure repeatability. Representative PBS loops are presented at five minutes of cycling, and indicate clockwise bubble expansion and counterclockwise compression (Figure 4). Average γ ± σ at selected time intervals are presented in Table II. Percent compression is defined here as 100*[(SAmax − SA20)/(SAmax)], where SAmax was the maximum SA value at expansion, and SA20 was the SA at which γ first reaches 20 mN m−1 upon compression (Figure 4, inset table).
We gratefully acknowledge Yoriel Marcano, Ann M. Czyzewski, and Michael Connolly, for their valuable assistance, and Mark Johnson for PBS use. This work was supported by the US National Institutes of Health (Grant 2 R01 HL67984) and the US National Science Foundation (Grant BES-0101195 and Collaborative Research in Chemistry Grant CHE-0404704). Portions of this work were performed at the Molecular Foundry, Lawrence Berkeley National Laboratory, which is supported by the Office of Science, Office of Basic Energy Sciences, of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.
SUPPLEMENTAL DATA. Details of azide and alkyne peptoid monomer synthesis, CD spectra, quantitative content in the lipid film, and tabulated PBS adsorption data. This material is available free of charge via the Internet at.