Design of α/β-Hybrid Peptide Ligands of α4β1 Integrin Equipped with a Linkable Side Chain for Chemoselective Biofunctionalization of Microstructured Materials

Arg-Gly-Asp (RGD)-binding integrins, e.g., αvβ3, αvβ1, αvβ5 integrins, are currently regarded as privileged targets for the delivery of diagnostic and theranostic agents, especially in cancer treatment. In contrast, scarce attention has been paid so far to the diagnostic opportunities promised by integrins that recognize other peptide motifs. In particular, α4β1 integrin is involved in inflammatory, allergic, and autoimmune diseases, therefore, it represents an interesting therapeutic target. Aiming at obtaining simple, highly stable ligands of α4β1 integrin, we designed hybrid α/β peptidomimetics carrying linkable side chains for the expedient functionalization of biomaterials, nano- and microparticles. We identified the prototypic ligands MPUPA-(R)-isoAsp(NHPr)-Gly-OH (12) and MPUPA-Dap(Ac)-Gly-OH (13) (MPUPA, methylphenylureaphenylacetic acid; Dap, 2,3-diamino propionic acid). Modification of 12 and 13 by introduction of flexible linkers at isoAsp or Dap gave 49 and 50, respectively, which allowed for coating with monolayers (ML) of flat zeolite crystals. The resulting peptide–zeolite MLs were able to capture selectively α4β1 integrin-expressing cells. In perspective, the α4β1 integrin ligands identified in this study can find applications for preparing biofunctionalized surfaces and diagnostic devices to control the progression of α4β1 integrin-correlated diseases.


Introduction
Integrins are one of the major families of adhesion receptors that mediate cell-cell and cell-extracellular matrix interactions (ECM). These heterodimeric transmembrane proteins are composed by diverse α and β subunits, giving rise in mammals to 24 possible α/β heterodimers [1]. Integrins, expressed on many cell types, regulate fundamental functions, such as adhesion, signaling, and viability, but they are also deeply involved in a variety of diseases, including the initiation and progression of cancer, coronary diseases, inflammatory, and autoimmune pathologies. Consequently, antibodies (Ab) or small molecules that block integrin functions have attracted significant attention as potential drugs [2].
On the other hand, very little attention has been paid so far to the use of integrin-binding motifs different from RGD for material biofunctionalization. In particular, α4β1 integrin recognizes the Leu-Asp-Val-Pro (LDVP) peptide in FN, the Leu-Asp-Thr-Ser (LDTS) sequence in the mucosal addressing cell adhesion molecule-1 (MAdCAM-1), and Ile-Asp-Ser (IDS) in VCAM-1. The α4β1 integrin, also known as CD49d/CD29 or as very late antigen-4 (VLA-4), is expressed on most leukocytes and plays a crucial role in inflammation, allergy, and autoimmune diseases, e.g., multiple sclerosis (MS), but also in cancer development, metastasis, and stem cell mobilization or retention [19]. Targeting α4β1 integrin with the humanized monoclonal antibody (mAb) natalizumab has already proven to be successful for the treatment of highly active relapsing and remitting MS [20]. Several peptidomimetic α4β1 integrin antagonists are currently under investigation for diverse pathologies, such as asthma, [21] allergic conjunctivitis, [22] age-related macular degeneration (AMD) [23], etc.
On the other hand, very little attention has been paid so far to the use of integrinbinding motifs different from RGD for material biofunctionalization. In particular, α4β1 integrin recognizes the Leu-Asp-Val-Pro (LDVP) peptide in FN, the Leu-Asp-Thr-Ser (LDTS) sequence in the mucosal addressing cell adhesion molecule-1 (MAdCAM-1), and Ile-Asp-Ser (IDS) in VCAM-1. The α4β1 integrin, also known as CD49d/CD29 or as very late antigen-4 (VLA-4), is expressed on most leukocytes and plays a crucial role in inflammation, allergy, and autoimmune diseases, e.g., multiple sclerosis (MS), but also in cancer development, metastasis, and stem cell mobilization or retention [19]. Targeting α4β1 integrin with the humanized monoclonal antibody (mAb) natalizumab has already proven to be successful for the treatment of highly active relapsing and remitting MS [20]. Several peptidomimetic α4β1 integrin antagonists are currently under investigation for diverse pathologies, such as asthma [21], allergic conjunctivitis [22], age-related macular degeneration (AMD) [23], etc.
In perspective, materials functionalized with α4β1 integrin-binding peptides might consent to design devices for the detection and quantification of leucocytes expressing active α4β1 integrins. The opportunity to detect leukocyte integrin binding for diagnostic applications has been proposed by several authors [24][25][26][27][28][29]. For instance, the binding of the monoclonal antibody N29 to activated β1 integrin on eosinophils was exploited in a the monoclonal antibody N29 to activated β1 integrin on eosinophils was explo patented kit related to asthma control; the integrin amount above a minimum th was used as an indicator of decreased lung function [30].
Within this field, very recently we designed a biomaterial constituted b monolayer of microcrystals, capable of reproducing the cell-adhesive multival tegrin-ligand interaction at the endothelial surfaces in the proximity of the si flammation [31]. The monolayers were coated with a peptide ligand derived f potent α4β1 integrin antagonist BIO1211 (Figure 1), MPUPA-Leu-Asp-Valwhich includes the LDV minimal epitope of the endogenous ligand FN, plus the moiety (methylphenylureaphenylacetic acid) at the N-terminus, which stro creased the affinity for α4 integrins [32]. This microstructured surface allowed sig and selective detection of α4β1 integrin-expressing cells. Unfortunately, the utility of this prototypic device is diminished by the poor stability of BIO1211 in logical conditions, e.g., heparinized blood, plasma and rat liver, lung and intes mogenates, where it is metabolized by hydrolytic cleavage of the terminal d moiety, giving a sequence much less active than the parent compound [33,34].
These results prompted us to design analogues of the hybrid α/β-peptid taining diverse β-amino acid cores, equipped with versatile functional groups f dient connection to linkers ( Figure 2) suitable for chemoselective functionalizati ocompatible materials. In the boxes, model β 3 -or β 2 -aminoacids, i.e., β 3 -Ala and β 2 -Ala, and β 3 -or β 2 -residues carrying side chain functional groups suitable for diverse linker connections, i.e., isoAsp and Dap, respectively, are shown. Functional groups exploitable for peptide conjugation are rendered in red; these are terminated with n-propylamine in 11 and 12, and with Ac in 13 and 14.
These results prompted us to design analogues of the hybrid α/β-peptide 6 containing diverse β-amino acid cores, equipped with versatile functional groups for expedient connection to linkers ( Figure 2) suitable for chemoselective functionalization of biocompatible materials.
The Fmoc peptidyl resin was deprotected by treatment with Pip/DMF (2 × 4 mL), as reported above. MPUPA (0.3 mmol) was introduced in the presence of HOBt/DCC (0.36 mmol each) as the activating agents, under the same conditions described above.
After the usual work up, the resin was washed twice with a mixture of THF/DMF/H 2 O (2:2:1 v/v/v) and subsequently swollen with the same mixture. After 30 min, a solution of PIFA (0.3 mmol) in THF/DMF/H 2 O (2:2:1) was added to the resin followed by pyridine (0.3 mmol) and the mixture was gently stirred at rt for 3 h. Then, the resin was washed with DMF (3 × 5 mL), DMF/DIPEA (9:1), and DMF (3 × 5 mL). The at least partially effective rearrangement was confirmed by a positive Kaiser test.
Cleavage from the resin was carried out using TFA/TIPS/H 2 O (80:10:10 v/v/v, 10 mL) at rt for 2.5 h. The reaction was filtered, the resin was washed with Et 2 O, and the mixture was concentrated at reduced pressure. The residue was transferred into a centrifugation vial and the peptide was precipitated with ice-cold Et 2 O. The crude peptide was purified by semipreparative RP HPLC (general methods) to give (S)-13 (70%, 96% pure) or (R)-14 (58%, 95% pure) as a white powder. Purity was assessed by analytical RP HPLC (general methods); Rt = 3.4 min. 1  MPUPA-Dap(Ahx)-Gly-OH.TFA 50. The MPUPA-(S)-Dap-Gly peptidyl resin 46 (0.2 mmol, estimated on the basis of resin load), prepared as described, was functionalized at the free α-amino group of Dap with Boc-Ahx-OH, preactivated with DCC/HOBt (0.5 mmol) in DMF (5 mL). Eventually, the Boc-protected peptidyl resin was treated with TFA/TIPS/H 2 O (80:10:10 v/v/v, 10 mL) at rt for 2.5 h. After the same work up procedure reported previously, the crude peptide was purified by semipreparative RP HPLC (general methods) to give (S)-15 (66%, 97% pure). Purity was assessed by analytical RP HPLC (general methods); Rt =1.9 min. 1 13

Cell Culture
All cell lines were purchased from the ATCC (Rockville, MD, USA). Jurkat E6.1 cells (expressing α4β1) were maintained as a stationary suspension culture in RPMI-1640 (Life Technologies, Carlsbad, CA, USA) supplemented with 2 mM glutamine and 10% FBS (fetal bovine serum). HEK293 cells were routinely cultured in EMEM (Lonza, Basel, Switzerland) enriched with 10% FBS, L-glutamine and nonessential amino acids. Cells were kept at 37 • C under a 5% CO 2 humidified atmosphere.
The number of adherent cells was obtained by comparison with a standard curve prepared in the same plate using diverse concentrations of cells. Experiments were carried out in quadruplicate and repeated at least three times. Data analysis and IC 50 values were calculated using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA).

Enzymatic Stability
Peptides 12, 13, and BIO1211, were dissolved in Tris buffer (pH 7.4) to give a concentration of 10 mM, and 10 µL aliquots of these stock solutions were added to 190 µL of mouse serum. Incubations were carried out at 37 • C. Aliquots of 20 µL were sampled over 180 min from the mixtures and the samples were diluted with 90 µL of glacial ACN for precipitating proteins. The mixtures were diluted with 90 µL of 0.5% AcOH and centrifuged (13,000× g, 15 min). The supernatants were collected and the amount of intact peptides was determined by RP HPLC. The experiments were performed in triplicate and repeated three times.

Preparation of the Biofunctionalized Surfaces 49-Zeolite MLs/50-Zeolite MLs, and of the Negative References BA-Zeolite MLs/49-Plates
A mixture of TEA (10 µL) and peptide 49 or 50 (1.5 mg) or n-butylamine (BA, 0.3 mg) in DMF (1 mL) was dissolved by sonication for 5 min. The IC-zeolite MLs (Supplementary Materials) were immersed in the solution and heated at 40 • C. After 3 h, the plates were removed and washed with DMF (5 mL) and EtOH (5 mL), then the MLs were air dried.
To prepare the negative control 49-plates, bare silica plates were reacted with ICPTES, and the resulting IC-plates were treated with peptide 49, under the same conditions utilized for the preparation of 49-zeolite MLs.

Cell Adhesion on 49-Zeolite MLs/50-Zeolite MLs/BA-Zeolite MLs/49-Plates
Jurkat or HEK293 (10 6 cells/mL) cells were suspended in 1% BSA in PBS in the presence of the fluorescent dye CMFDA (6.25 µM) and incubated for 20 min at 37 • C while gently shaken. At the end of the incubation, the cells were washed three times with 1 mL of 1% BSA in PBS.
The peptide MLs were washed with PBS solution, then they were placed in a 6-well plate (growth area 8.87 cm 2 ) and seeded with 5 × 10 5 Jurkat or HEK-293 cells. Each well was filled with PBS (final volume, 1 mL), and the plates were incubated for 15 min. Afterwards, each well was washed twice with PBS and the sample fixed with 3% PFA for 10 min at RT. The plates were washed twice with 0.1% glycine in PBS, and each monolayer was embedded in Mowiol and analyzed with a confocal laser-scanning microscope.

Confocal Microscopy
For visualization of the DXP-loaded zeolites in red, fluorescence was excited with a λ = 543 nm laser and detected at 650 nm; CMFDA-stained cells were visualized in green by excitation at 488 nm and detected at λ = 530 nm. Images were quantified by ImageJ software version 1.52 t, NIH, USA. The number of cells adherent to a surface of 0.0025 cm 2 was determined three times, and the average was expressed as cell.cm −2 ± S.D. Statistical comparison was performed using a Student's t-test; p < 0.05 was considered significant.

Synthesis of β-Amino Acids
Many β-amino acids are commercially available in diverse protected forms theless, we opted for in-house preparations, potentially exploitable for preparin ety of (S)-or (R)-, natural or unnatural β-residues. β 3 -amino acids can be obtaine rect Arndt-Eistert homologation of the corresponding α-amino acids. In sever this procedure was found to be poorly efficient and not suitable for large scale p tion [42]. Therefore, we utilized the scalable procedure reported by Caputo et. al.
The β-amino alcohol (S)-15 or (R)-16 was obtained from (S)-or (R)-Boc-Ala spectively, by reduction of the mixed anhydride with NaBH4. The β-amino alco then converted into the corresponding β-amino iodide (S)-17 or (R)-18 by mean phenylphosphine/iodide complex. Due to poor stability under the common separation and purification procedu iodoamino intermediate was used directly without isolation. The subsequent d ment of iodine with potassium cyanide led to (S)-19 or (R)-20, and the final hy gave Boc-β 3 -homoAla-OH (S)-21 or (R)-22 (Scheme 2). The risk of racemization cluded by HPLC on a chiral column and polarimetry, as reported [43].
The enantiopure model residue (S)-or (R)-β 2 -Ala was synthesized as methyl a modified version of a multi-step procedure previously reported by Lee et al sylation of (S)-or (R)-β-hydroxy ester provided (S)-23 or (R)-24 in good yield, 8 91%, respectively (Scheme 3). The tosylate was treated with N N,N-dimethylformamide (DMF) to afford the corresponding azide (S)-25 or (R nally, the catalytic reduction of the crude azide with H2 and the catalytic amount gave the primary amine (S)-27 or (R)-28.

Synthesis of β-Amino Acids
Many β-amino acids are commercially available in diverse protected forms. Nevertheless, we opted for in-house preparations, potentially exploitable for preparing a variety of (S)-or (R)-, natural or unnatural β-residues. β 3 -amino acids can be obtained by direct Arndt-Eistert homologation of the corresponding α-amino acids. In several cases, this procedure was found to be poorly efficient and not suitable for large scale preparation [42]. Therefore, we utilized the scalable procedure reported by Caputo et. al. (Scheme 2) [43].

Synthesis of β-Amino Acids
Many β-amino acids are commercially available in diverse protected forms. Nevertheless, we opted for in-house preparations, potentially exploitable for preparing a variety of (S)-or (R)-, natural or unnatural β-residues. β 3 -amino acids can be obtained by direct Arndt-Eistert homologation of the corresponding α-amino acids. In several cases, this procedure was found to be poorly efficient and not suitable for large scale preparation [42]. Therefore, we utilized the scalable procedure reported by Caputo et. al. (Scheme 2) [43].
The β-amino alcohol (S)-15 or (R)-16 was obtained from (S)-or (R)-Boc-Ala-OH, respectively, by reduction of the mixed anhydride with NaBH4. The β-amino alcohol was then converted into the corresponding β-amino iodide (S)-17 or (R)-18 by means of triphenylphosphine/iodide complex. Due to poor stability under the common separation and purification procedures, the iodoamino intermediate was used directly without isolation. The subsequent displacement of iodine with potassium cyanide led to (S)-19 or (R)-20, and the final hydrolysis gave Boc-β 3 -homoAla-OH (S)-21 or (R)-22 (Scheme 2). The risk of racemization was excluded by HPLC on a chiral column and polarimetry, as reported [43].
The enantiopure model residue (S)-or (R)-β 2 -Ala was synthesized as methyl ester by a modified version of a multi-step procedure previously reported by Lee et al [44]. Tosylation of (S)-or (R)-β-hydroxy ester provided (S)-23 or (R)-24 in good yield, 87% and 91%, respectively (Scheme 3). The tosylate was treated with NaN3 in N,N-dimethylformamide (DMF) to afford the corresponding azide (S)-25 or (R)-26. Finally, the catalytic reduction of the crude azide with H2 and the catalytic amount of Pd/C gave the primary amine (S)-27 or (R)-28. The β-amino alcohol (S)-15 or (R)-16 was obtained from (S)-or (R)-Boc-Ala-OH, respectively, by reduction of the mixed anhydride with NaBH 4 . The β-amino alcohol was then converted into the corresponding β-amino iodide (S)-17 or (R)-18 by means of triphenylphosphine/iodide complex.
Due to poor stability under the common separation and purification procedures, the iodoamino intermediate was used directly without isolation. The subsequent displacement of iodine with potassium cyanide led to (S)-19 or (R)-20, and the final hydrolysis gave Boc-β 3 -homoAla-OH (S)-21 or (R)-22 (Scheme 2). The risk of racemization was excluded by HPLC on a chiral column and polarimetry, as reported [43].
The enantiopure model residue (S)-or (R)-β 2 -Ala was synthesized as methyl ester by a modified version of a multi-step procedure previously reported by Lee et al [44]. Tosylation of (S)-or (R)-β-hydroxy ester provided (S)-23 or (R)-24 in good yield, 87% and 91%, respectively (Scheme 3). The tosylate was treated with NaN 3 in N,N-dimethylformamide (DMF) to afford the corresponding azide (S)-25 or (R)-26. Finally, the catalytic reduction of the crude azide with H 2 and the catalytic amount of Pd/C gave the primary amine (S)-27 or (R)-28.  Finally, we selected a β 2 -amino acid core carrying an amino group side chain, i.e., (S)-or (R)-2,3-diaminopropionic acid (Dap) (Figure 2). For the preparation of this residue, we opted for the Hofmann rearrangement of N-protected (S)-or (R)-Asn in solid phase, giving the corresponding Dap peptidyl resin (Scheme 8). In this case, the amide side chain of Asn was intended as a convenient protected precursor of the β-amino group of Dap, orthogonal to Fmoc.

Synthesis of Hybrid Peptides
The hybrid sequences 7 and 8 were prepared in solution. Boc-β 3 -Ala (S)-21 or (R)-22 was coupled with H-Gly-OBn . TFA in the presence of the coupling agents EDC/HOBt/TEA. The resulting Boc-dipeptide 33 or 34 was treated with TFA and the deprotected dipeptide 35 or 36 was N-capped with MPUPA using the same activating agents. Catalytic hydrogenation of the benzyl ester 37 or 38 afforded (S)-7 or (R)-8 in good yield after purification by semipreparative reversed phase (RP)-HPLC (7, 76%; 8, 87%; Scheme 5).   Finally, deprotection of the tBu group with TFA gave the final peptidomimetic (S)-9 or (R)-10 in almost quantitative yield (Scheme 6) and excellent purity (95 and 94%) after precipitation from ether.
Finally, deprotection of the tBu group with TFA gave the final peptidomimetic (S)-9 or (R)-10 in almost quantitative yield (Scheme 6) and excellent purity (95 and 94%) after precipitation from ether.  Finally, (S)-13 or (R)-14 were assembled by SPPS on a Wang resin preloaded with Fmoc-Gly (Scheme 8). After removal of Fmoc group with 20% Pip/DMF, Fmoc protected (S)-or (R)-Asn-OH was activated for coupling with (S)-or (R)-Fmoc-Asn-OH in the presence of DCC/HOBt. The Hofmann rearrangement of the primary amide of Asn into the corresponding amine was performed on resin using PhI(OCOCF 3 ) 2 (PIFA) and pyridine in DMF/H 2 O [45]. The resulting Nα-Fmoc-Dap residue was coupled at the β-amino group with MPUPA under the same conditions described above. dine in DMF/H2O [45]. The resulting Nα-Fmoc-Dap residue was coupled at the β-amino group with MPUPA under the same conditions described above.
Then, Fmoc deprotection of Dap gave the peptidyl resin (S)-46 or (R)-47. The subsequent acetylation with Ac2O and pyridine completed the linear sequence. Peptide cleavage was accomplished by treatment of the resin with TFA and scavengers (Scheme 8). The crude product 13 or 14 was precipitated from ice-cold Et2O, collected by centrifuge, and purified by semipreparative RP-HPLC (general methods).
In general, after synthesis the crude peptides of all sequences were purified by semipreparative RP HPLC over a C18 column (general methods), using H2O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was assessed to be >95% by RP HPLC; chemical identity was confirmed by ESI-MS, 1 H, 13 C and 2D gCosy NMR spectroscopy.

Potency and Selectivity of Hybrid Peptides for Human α4β1 Integrin as Measured by Inhibition of Integrin-Mediated Cell Adhesion to Endogenous Ligands
The ability of the reference compounds BIO1211 and 6, and of the new hybrid peptides 7-14, to modulate integrin-mediated cell adhesion to the α4β1 endogenous ligand FN, was measured by cell adhesion assays. Integrin ligands able to inhibit cell adhesion promoted by the endogenous ligand were referred to as antagonist. Jurkat cells, an immortalized cell line of human T lymphocytes, are often utilized as prototypic α4β1 integrin-expressing cells [22,46]. The reference compound BIO1211 and the hybrid peptide 6 prevented cell adhesion with nanomolar IC50 values (Table 1), as previously described [22,32]. In general, after synthesis the crude peptides of all sequences were purified by semipreparative RP HPLC over a C18 column (general methods), using H 2 O/acetonitrile mixtures containing 0.1% trifluoroacetic acid; purity was assessed to be >95% by RP HPLC; chemical identity was confirmed by ESI-MS, 1 H, 13 C and 2D gCosy NMR spectroscopy.

Potency and Selectivity of Hybrid Peptides for Human α4β1 Integrin as Measured by Inhibition of Integrin-Mediated Cell Adhesion to Endogenous Ligands
The ability of the reference compounds BIO1211 and 6, and of the new hybrid peptides 7-14, to modulate integrin-mediated cell adhesion to the α4β1 endogenous ligand FN, was measured by cell adhesion assays. Integrin ligands able to inhibit cell adhesion promoted by the endogenous ligand were referred to as antagonist. Jurkat cells, an immortalized cell line of human T lymphocytes, are often utilized as prototypic α4β1 integrin-expressing cells [22,46]. The reference compound BIO1211 and the hybrid peptide 6 prevented cell adhesion with nanomolar IC 50 values (Table 1), as previously described [22,32]. BIO1211-tides 7-14, to modulate integrin-mediated cell adhesion to the α4β1 endogenous ligand FN, was measured by cell adhesion assays. Integrin ligands able to inhibit cell adhesion promoted by the endogenous ligand were referred to as antagonist. Jurkat cells, an immortalized cell line of human T lymphocytes, are often utilized as prototypic α4β1 integrin-expressing cells [22,46]. The reference compound BIO1211 and the hybrid peptide 6 prevented cell adhesion with nanomolar IC50 values (Table 1), as previously described [22,32]. The ability of the reference compounds BIO1211 and 6, and of the new hybrid peptides 7-14, to modulate integrin-mediated cell adhesion to the α4β1 endogenous ligand FN, was measured by cell adhesion assays. Integrin ligands able to inhibit cell adhesion promoted by the endogenous ligand were referred to as antagonist. Jurkat cells, an immortalized cell line of human T lymphocytes, are often utilized as prototypic α4β1 integrin-expressing cells [22,46]. The reference compound BIO1211 and the hybrid peptide 6 prevented cell adhesion with nanomolar IC50 values (Table 1), as previously described [22,32]. a In a cell-based assay, the adhesion of Jurkat cells to an immobilized endogenou ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are expressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC50 = 9.8 nM vs. IC50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain below the molecular plane (Figure 2 a In a cell-based assay, the adhesion of Jurkat cells to an immobilized endogenous ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are expressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC50 = 9.8 nM vs. IC50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain below the molecular plane (Figure 2 a In a cell-based assay, the adhesion of Jurkat cells to an immobilized endogenous ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are expressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC50 = 9.8 nM vs. IC50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain be- a In a cell-based assay, the adhesion of Jurkat cells to an immobilized endogenous ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are expressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC50 = 9.8 nM vs. IC50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain be- a In a cell-based assay, the adhesion of Jurkat cells to an immobilized endogenous ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are expressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC50 = 9.8 nM vs. IC50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain be- a In a cell-based assay, the adhesion of Jurkat cells to an immobilized endogenous ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are expressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central β-residue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC50 (4.06 μM), and the enantiomer 8 was defined as not active (i.e., IC50 > 5 μM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC50 = 9.8 nM vs. IC50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain be- a In a cell-based assay, the adhesion of Jurkat ells to an immobilized endogenous ligand (FN, 10 μg/mL) was measured. b Three independent experiments were run in quadruplicate. Data are ex-pressed as means ± SD. c [22].
Concerning the hybrid peptidomimetics, the IC 50 values strongly varied depending on the type of β-residue, on (S)-or (R)-configuration, as well as on the nature of the central βresidue side chain. Compound 7, containing (S)-β 3 -Ala, showed a very modest micromolar IC 50 (4.06 µM), and the enantiomer 8 was defined as not active (i.e., IC 50 > 5 µM). As for the peptides containing β 2 -Ala, both enantiomers (S)-9 and (R)-10 were not able to modulate α4β1 integrin-mediated cell adhesion to a significant extent. This result was somewhat unexpected, since the parent peptide DS-70 (6), containing (S)-β 2 -Pro, had a noteworthy nanomolar, IC 50 (4.3 ± 1.7 nM). Plausibly, the rigid structure of the pyrrolidine scaffold conferred to the entire peptide a constrained, more appropriate bioactive conformation [47].
The experiments performed on the remaining α/β hybrid peptides 11-14 gave highly contrasting results as compared to the model peptides 7-10 containing β 3 -or β 2 -Ala (Table 1). Peptide 12, containing the (R)-isoAsp(NHPr), showed much higher potency with respect to the (S) enantiomer 11, IC 50 = 9.8 nM vs. IC 50 > 5000 nM, respectively. Topologically, the (R)-configured β 3 -amino acid isoAsp is consistent with (S)-β 3 -Ala, since the side chains of both residues, when observed in the fully extended conformation, remain below the molecular plane ( Figure 2). On the other hand, peptide 13 containing the β 2 -residue (S)-Dap showed a meaningful IC 50 in the submicromolar range, while the enantiomer 14 was inactive. Apparently, the amide groups at the side chains of 12 and 13 strongly improved the compound's ability to modulate cell adhesion mediated by α4β1 integrin, since the topologically corresponding model peptides carrying a simple methyl side chain were practically inactive (Table 1).
These experiments pointed at 12 and 13 as effective antagonist ligands of α4β1 integrin, potentially useful for therapeutic applications in inflammatory and allergic pathologies. These opportunities will be fully developed in due course. In addition, for the presence of functionalized side chains, 12 and 13 can be conveniently exploited in the design and preparation of biomaterials capable of reacting to α4β1 integrin cells, provided that these compounds display sufficient stability in physiological conditions.

In Vitro Enzymatic Stability of 12, 13
The stability of the hybrid peptides MPUPA-(R)-isoAsp(NHPr)-Gly-OH (12) and MPUPA-(S)-Dap(Ac)-Gly-OH (13) was estimated in vitro by incubation in mouse serum at 37 • C, in comparison with BIO1211. The incubation mixture was sampled during 180 min and the quantity of intact peptides was determined by HPLC analysis. As shown in the Supplementary Materials, BIO1211 was rapidly hydrolyzed, since only about 10% of the initial amount was still present after 1 h. On the other hand, the peptides 12 and 13 appeared to be much more stable under the same conditions, and after 3 h the amount of the two peptides was around 80% and 91%, respectively. These data provide evidence that the presence of the β-amino acid markedly increased the enzymatic stability.
The comparatively higher stability of 13 vs. 12 was not completely unexpected. Indeed, the (S)-β 2 -amino acid Dap is topologically consistent with an (R)-α-amino acid ( Figure 2). Hence, the higher stability of 13 could correlate with the presence of the β-residue as well as with the unnatural display of its side chain.

Design of Linkable α4β1 Integrin Ligands
The α4β1 integrin ligands MPUPA-(R)-isoAsp(NHPr)-Gly-OH (12) and MPUPA-Dap(Ac)-Gly-OH (13) were selected for the biofunctionalization of microstructured monolayers to provide surfaces potentially capable of recognizing and binding α4β1 integrinexpressing cells. Although the hybrid peptide 12 was by far the most potent integrin ligand (Table 1), we utilized 13 also, since the latter appeared to be slightly more stable, as estimated by incubating the peptides in mouse serum (see above). Hence, we modified the structures of 12 and 13 (Schemes 9 and 10) by introduction of effective linkers for anchoring onto the surfaces via urea linkages [17]).
For the modification of peptide 12, (R)-Fmoc-Asp(OBn)-OH was reacted with Boc monoprotected 1,6-hexanediamine (Boc-Hda, Supplementary Materials) using EDC/HOBt/TEA as activating agents. Hydrogenolysis gave Fmoc-isoAsp(Boc-Hda)-OH 48 (Scheme 9). Subsequently, 48 was utilized for the SPPS of the peptide MPUPA-(R)-isoAsp(Hda)-Gly-OH.TFA (49), under the same conditions described for the preparation of 12. The treatment with TFA in the presence of scavengers allowed at the same time peptide cleavage and the removal of the Boc group at Hda (Scheme 9). Peptide 13 was modified by substituting the acetyl group with 6-aminohexanoic acid (Ahx), giving the peptide MPUPA-Dap(Ahx)-Gly-OH.TFA (50). In detail, the MPUPA-(S)-Dap-Gly peptidyl resin 46, prepared as described in Scheme 8, was functionalized at the free α-amino group with Boc-Ahx-OH, activated with DCC/HOBt. Eventually, the treatment of the Boc-protected peptidyl resin with TFA/scavengers allowed to cleave the peptide and to deprotect the amino group of Ahx (Scheme 10).
For the modification of peptide 12, (R)-Fmoc-Asp(OBn)-OH was reacted with Boc mono-protected 1,6-hexanediamine (Boc-Hda, Supplementary Materials) using EDC/HOBt/TEA as activating agents. Hydrogenolysis gave Fmoc-isoAsp(Boc-Hda)-OH 48 (Scheme 9). Subsequently, 48 was utilized for the SPPS of the peptide MPU-PA-(R)-isoAsp(Hda)-Gly-OH . TFA (49), under the same conditions described for the preparation of 12. The treatment with TFA in the presence of scavengers allowed at the same time peptide cleavage and the removal of the Boc group at Hda (Scheme 9).

Discussion
Most inflammatory diseases, including allergies and asthma, are characterized by α4β1 integrin-mediated extravasation and exaggerated accumulation of leukocytes [52]. Indeed, the activation of lymphocytes promotes both integrin clustering and their switch to a higher affinity conformation, and these events rapidly strengthen cell interaction with the endogenous ligands overexpressed on the vascular surface of activated endothelial tissue [19]. Leukocyte migration is also typical of autoimmune diseases characterized by chronic inflammation, e.g., rheumatoid arthritis, autoimmune encephalomyelitis, and multiple sclerosis [53].
In this scenario, the development of materials capable of interacting specifically with lymphocytes by specific recognition of α4β1 integrins could find practical applications in diagnostics (see Introduction). As a proof of concept, herein we describe the preparation of monolayers of zeolites coated with novel ligands inspired from the potent α4β1 integrin antagonist BIO1211 [32], stably bonded to the crystals via urea linkages. Generally, the practicability of peptide-material conjugates is questioned mainly because the pep- v/v/v/v), rt, 2.5 h. Peptide 13 was modified by substituting the acetyl group with 6-aminohexanoic acid (Ahx), giving the peptide MPUPA-Dap(Ahx)-Gly-OH . TFA (50). In detail, the MPUPA-(S)-Dap-Gly peptidyl resin 46, prepared as described in Scheme 8, was functionalized at the free α-amino group with Boc-Ahx-OH, activated with DCC/HOBt. Eventually, the treatment of the Boc-protected peptidyl resin with TFA/scavengers allowed to cleave the peptide and to deprotect the amino group of Ahx (Scheme 10).
To immobilize the resulting isocyano (IC)-zeolites, the AP-plates were immersed in a suspension of IC-zeolites, giving IC-zeolite monolayers (IC-zeolite MLs, Figure 3a), connected by highly stable urea linkages (Scheme 11). The efficacy of each subsequent functionalization of the native zeolites with organic molecules was assessed by the increase in the percentage values of C(1s) and N(1s) as determined by X-ray photoelectron spectrometry (XPS) analysis (Table 2 and Supplementary Materials). In particular, the IC-zeolite MLs showed C(1s) and N(1s) values (%) similar to the values obtained for the unbound zeolites. Thermogravimetric analyses (TGA) verified a higher weight loss for the functionalized zeolites as compared to the native zeolites (Supplementary Materials). Eventually, the IC-zeolite MLs were immersed in a solution of peptide 49 or 50 in DMF in the presence of TEA to give the monolayers 49-zeolite MLs or 50-zeolite MLs (Scheme 11). XPS analysis confirmed the large increase of C(1s)/N(1s) percentages.
To challenge the stability under physiological conditions prior to performing the cell adhesion experiments, peptide-zeolite MLs were immersed at rt for 24 h in mouse serum diluted in sterile milli-Q water. Then, the monolayers were washed with distilled water. Subsequent XPS analysis showed the same atomic percent values, confirming the stability of the peptidic functionalization (not shown).
Alternatively, IC-zeolite MLs were functionalized with n-butylamine (BA) to afford BA-zeolite MLs, designed as peptide-negative references. XPS analysis showed a inferior increase of C(1s)/N(1s) for BA-zeolite MLs as compared to peptide-zeolite MLs (Table 2 and Supplementary Materials).
As a further control, silica plates were directly functionalized with peptide 49 without interposing the crystals (see Discussion). To the purpose, bare silica plates were reacted with ICPTES, and the IC-plates were treated with peptide 49, under the same conditions utilized for the preparation of 49-zeolite MLs. XPS analysis of the resulting 49-plates was suggestive of a moderate functionalization (Table 2 and Supplementary Materials).

Cell Adhesion Experiments
To measure cell adhesion to peptidyl-zeolite MLs, experiments were performed by seeding either Jurkat cells, mainly expressing α4β1 integrin [19,31], or HEK-293 control cells not expressing α4 integrin, which are not expected to adhere upon rapid incubation [51]. Cells were stained with a green fluorescent dye (CMFDA) and seeded on the peptide monolayers. After a very brief incubation of 15 min, the non-adherent cells were washed away, while the attached cells were fixed with paraformaldehyde.

Discussion
Most inflammatory diseases, including allergies and asthma, are characterized by α4β1 integrin-mediated extravasation and exaggerated accumulation of leukocytes [52]. Indeed, the activation of lymphocytes promotes both integrin clustering and their switch to a higher affinity conformation, and these events rapidly strengthen cell interaction with the endogenous ligands overexpressed on the vascular surface of activated endothelial tissue [19]. Leukocyte migration is also typical of autoimmune diseases characterized by chronic inflammation, e.g., rheumatoid arthritis, autoimmune encephalomyelitis, and multiple sclerosis [53].
In this scenario, the development of materials capable of interacting specifically with lymphocytes by specific recognition of α4β1 integrins could find practical applications in diagnostics (see Introduction). As a proof of concept, herein we describe the preparation of monolayers of zeolites coated with novel ligands inspired from the potent α4β1 integrin antagonist BIO1211 [32], stably bonded to the crystals via urea linkages. Generally, the practicability of peptide-material conjugates is questioned mainly because the peptides grafted onto biomaterial surfaces might be rapidly hydrolyzed by serum proteins in vivo, thus nullifying the functionalization [9,10,54,55]. In point of fact, BIO1211 was found to have poor stability under physiological conditions [33,34]. As a consequence, we took into consideration the use of more stable peptidomimetic ligands. Of notice, there are at present no stable ligands available to specifically bind to α4β1 integrin while simultaneously carrying suitable linkers for bioconjugation. This is in net contrast to the class of RGDbinding integrins, for which a variety of suitable ligands has been proposed (Figure 1), including the well-known cyclopeptide c(RGDfK) (1) [15], conveniently equipped with a flexible aminoalkyl side chain.
The ability of these compounds to modulate α4β1 integrin-mediated cell adhesion was assayed in vitro in comparison with the reference compound BIO1211. The best results were observed for MPUPA-(R)-isoAsp(NHPr)-Gly-OH (12) and MPUPA-Dap(Ac)-Gly-OH (13) ( Table 1). The comparison between these peptides and the model peptides containing β-Ala suggested that the presence of amide groups at the side chains of 12 and 13 may significantly contribute to a compound's ability to inhibit α4β1-mediated cell adhesion. The detailed study of the effects exerted by the different substituents on ligand-receptor interactions is beyond the scope of this work. Further studies by molecular modeling and molecular docking are currently in progress in our group.
For the presence of functionalized side chains, which can be equipped with taggable or linkable side chains, 12 and 13 were selected for the preparation of biomaterials. To the purpose, the peptides were modified by introducing suitable linkers, giving MPUPA-(S)-isoAsp(Hda)-Gly-OH (49) and MPUPA-(S)-Dap(Ahx)-Gly-OH (50), respectively (Schemes 9 and 10). Thereafter, the monolayers 49-zeolite ML and 50-zeolite MLs were expediently prepared by stably anchoring the peptides onto MLs of zeolite L crystals mounted over silica plates (Scheme 10).
α4β1 Expressing Jurkat cells or integrin-negative HEK-293 cells were seeded onto the peptide MLs. Consistent with the high affinity of the parent 12 for α4β1 integrin, 49-zeolite MLs showed the highest number of adherent Jurkat cells, as revealed by confocal microscopy analysis. Both peptide MLs showed an excellent discrimination of Jurkat cells with respect to reference HEK-293 cells (Figure 3g). The latter are known to express significant levels of β1 integrin but very low levels of α4 integrin [31,51]. As expected, Jurkat cells showed negligible adhesion onto MLs deprived of the peptide (BA-zeolite ML).
Interestingly enough, Jurkat cells showed little adhesion to monolayers of peptides directly anchored to the silica plates (49-plates). Plausibly, the nanostructured surface of zeolite L crystals significantly fosters cell adhesion. Due to their flat shape and peculiar microstructure, during the formation of the monolayers the nanocrystals tend to adopt a clear orientation, with the cavities perpendicular to the support surface (Scheme 11) [56]. As a consequence, the paved surface appears interspersed with a regular alternation of solid parts and cavities, so that the bonded ligands maintain a regular spacing from each other, forming a well-organized biofunctionalized surface. Integrin-mediated adhesion is known to be regulated by multiple features of the adhesive surface, including topography and spacing between the ligands [57]. Plausibly, the resulting nanostructured surface might enhance cell recognition due to the regular positioning of the biomolecules, correlated in turn to the regular distance between adjacent channels in the nanocrystals.

Conclusions
On the basis of the results above discussed, zeolite MLs, functionalized with hybrid α/β-peptidomimetic ligands of α4β1 integrin, may be exploited for preparing functionalized biomaterials for a wide range of medical applications, from therapy to diagnosis, including the possibility of developing diagnostic devices capable of detecting desired cells in biological fluids obtainable from patients in a noninvasive way. In perspective, more functional assays will be performed to better investigate the effect of the hybrid ligands on transmigration, chemotaxis, and intracellular signaling in inflammatory cells, including neutrophils and monocytes. Moreover, the peptide-coated materials might be utilized as model systems to study in detail the mechanisms of integrin-mediated adhesion and signaling using different types of inflammatory cells.