Chemically Synthesized SDF-1α Analogue, N33A, Is a Potent Chemotactic Agent for CXCR4/Fusin/LESTR-expressing Human Leukocytes*

Stromal cell-derived factor (SDF) 1 is a potent chemoattractant for leukocytes through activation of the receptor CXCR4/Fusin/LESTR, which is a fusion co-factor for the entry of T lymphocytotropic human immunodeficiency virus type 1 (HIV-1). This CXCR4-mediated HIV-1 fusion can be inhibited by SDF-1. Because of its importance in the study of immunity and AIDS, large scale production of SDF-1 is desirable. In addition to recombinant technology, chemical synthesis provides means by which biologically active proteins can be produced not only in large quantity but also with a variety of designed modifications. In this study, we investigated the binding and function of an SDF-1α analogue, N33A, synthesized by a newly developed native chemical ligation approach. Radioiodinated N33A showed high affinity binding to human monocytes, T lymphocytes, as well as neutrophils, and competed equally well with native recombinant SDF-1α for binding sites on leukocytes. N33A also showed equally potent chemoattractant activity as native recombinant SDF-1α for human leukocytes. Further study with CXCR4/Fusin/LESTR transfected HEK 293 cells showed that N33A binds and induces directional migration of these cells in vitro. These results demonstrate that the chemically synthesized SDF-1α analogue, N33A, which can be produced rapidly in large quantity, possesses the same capacity as native SDF-1α to activate CXCR4-expressing cells and will provide a valuable agent for research on the host immune response and AIDS.

Stromal cell-derived factor (SDF) 1 1 has been reported to be a primordial chemokine of the CXC subfamily and has multiple biological activities on a variety of cell types (1)(2)(3)(4). SDF-1 was initially isolated as a T lymphocyte chemoattractant and was found to be active also on monocytes but not on neutrophils (1,3). However, its activity on neutrophils was subsequently shown by using Ca 2ϩ mobilization experiments (4). The SDF-1 gene is located on chromosome 10 (5), while the genes for other known CXC chemokines are located chromosome 4 (6,7). SDF-1 is well conserved with only a single amino acid substitution between human and murirne molecules (1,2). SDF-1␣ and SDF-1␤ were believed to be the result of differential splicing of a single gene, with SDF-1␣ missing four amino acids in the carboxyl terminus. The CXCR4/Fusin/LESTR has been identified as one of the functional receptors for SDF-1 (3,4), a member of the seven transmembrane spanning receptor superfamily (8,9). CXCR4 was initially cloned as an orphan receptor (10 -12) and was later identified as a fusion co-factor for the entry of HIV-1 of the T lymphotropic strain (13). In addition to its activities on human leukocytes, SDF-1 has been shown to inhibit the fusion and replication of T lymphotropic HIV-1 in host cells bearing CD4 and CXCR4 (3,4). Therefore, SDF-1 plays a pivotal role in host immune system and its defense against infection.
Rapid production of cytokines and chemokines is essential for structure-function studies and design of molecules with agonist or antagonist activities. The turbocharged peptide synthesis technology (14) and chemical ligation of peptide segments (15) permit the synthesis of chemokines in large quantity. This approach was therefore utilized to synthesize SDF-1␣ based on its great utility in studies of immune responses and infectious diseases. However, the thioacid peptide (SDF-1␣(1-33)-␣COSH) could not be readily generated on the standard ␣-thiocarboxylate resin (16) due to the presence of an Asn 33 residue at the ligation site in SDF-1␣. Instead, alanine was coupled to the ␣-thiocarboxylate resin, then the mutant SDF-1␣(1-33, N33A) was synthesized. In this study, we report that this SDF-1␣ analogue, N33A, displays full biological activity through the activation of the receptor CXCR4. This analogue, which can be rapidly produced in large quantity in the absence of contaminating peptides, will prove to be an important tool in the study of host immunity and AIDS.

MATERIALS AND METHODS
Chemokines-Human recombinant SDF-1␣ was purchased from Pep-roTech Inc. (Rocky Hill, NJ). Radioiodination of the SDF-1␣ and the analogue N33A was performed by a lactoperoxidase-labeling procedure. The radioactive ligands were further purified by reversed phase HPLC. The specific activity of the radioiodinated chemokines was 2200 Ci/mmol.
Chemical Synthesis of SDF-1␣ Analogue, N33A-Boc protected amino acids were obtained from the following sources: AnaSpec (San Jose, CA), Bachem (Philadelphia, PA), NovaBiochem (San Diego, CA), and Peptides International (Louisville, KY). Peptides were synthesized on a modified ABI430A instrument using in situ neutralization Boc chemistry protocols (17). C-terminal segments were prepared on -OCH2Pam resins (ABI, Foster City, CA). N-terminal segments were prepared on ␣-thiocarboxylate resin (16). Standard HF cleavage protocols were employed following N-terminal Boc removal and drying of the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ Ligation of the peptide segments was performed at 4 mM peptide concentration in 6 M guanidine, 0.1 M Tris, pH 7, in the presence of 33 mM thiophenol (Fluka, Switzerland) at room temperature. Ligation was monitored by HPLC and was typically complete within 24 h. Ligation was followed by HPLC purification and lyophilization. After purification, the full-length peptide was reduced at 1 mg/ml in 8 M urea (Fluka), 0.1 M Tris (Fluka), 5.37 mM EDTA ((Fluka), pH 8.6, in the presence of 100 mM 2-mercaptoethanol (Fluka). Reduction occurs under a nitrogen atmosphere at 40°C for 1 h. After complete reduction, the mixture was reconstituted with the same buffer at 0.2 mg/ml with 18.7 mM oxidized glutathione (Sigma). The solution was dispensed into a Spectrum Spectra/Por*7 dialysis membrane (Houston, TX) (M r cut-off, 3500) and the bag was placed in 1 liter of initial dialysis buffer of 8 M urea, 0.1 M Tris, 1 mM EDTA, 3 mM 2-mercaptoethanol, 1.3 mM oxidized glutathione, pH 8.6. Over a period of 2 days, 4 liters of 2 M urea, 0.1 M Tris, pH 8.6, was pumped into the vessel containing the dialysis bag (18). After lyophilization, the full-length peptide was oxidized at 1 mg/ml in 2 M guanidine HCl (Fluka), 0.1 M Tris (Fluka), pH 8.6, at room temperature in the presence of air. Folding was complete after stirring overnight and was monitored by HPLC and mass spectrometry.
Leukocyte migration was evaluated using a 48-well microchamber (Neuroprobe, Cabin John, MD) technique as described previously (19,20,22). The migration of CXCR4/293 cells was also assessed by the 48-well microchamber technique with the polycarbonate filters (10-m pore size) (21) precoated with collagen type I (Collaborative Biomedical Products, Bedford, MA). The results are expressed as the the chemotaxis index (CI) representing the fold increase in the cell migration induced by stimuli versus control medium. All experiments were performed at least two times, and results from one experiment are shown. The statistical significance of the difference between migration in re-sponse to stimuli and control was assessed by Student's t test.
Binding Assays with Radiolabeled SDF-1␣ and N33A-Binding assays were performed using a single concentration of 125 I-labeled chemokines in the presence of increasing concentrations of unlabeled ligands (19,21). The binding data were analyzed with a Macintosh computer program LIGAND (P. Munson, Division of Computer Research and Technology, NIH, Bethesda, MD). In Scatchard plots, the binding data were analyzed with both "one-site" and "two-site" models, and only the one-site model better fit the curves obtained with either native leukocytes or CXCR4/293 cells. The rate of competition for binding by unlabeled ligands was calculated with the following formula: % inhibition ϭ 1 Ϫ (binding in the presence of unlabeled chemokine/binding in the presence of medium alone) ϫ 100 (Eq. 1)

RESULTS AND DISCUSSION
We first examined whether the SDF-1␣ analogue N33A was able to bind and activate human peripheral blood monocytes and T cells. Fig. 1 shows that 125 I-N33A specifically bound to human peripheral blood monocytes (Fig. 1A) and T lymphocytes ( Fig. 1B) with high affinity (1.8 and 1.4 nM for monocytes and T cells, respectively). This level of binding to monocytes and T cells by 125 I-N33A was comparable to that of 125 I-rhSDF-1␣ as examined in parallel experiments (data not shown). Unlabeled N33A displaced 125 I-rhSDF-1␣ binding to both monocytes and T lymphocytes (Fig. 1, C and D), and likewise, the binding of 125 I-N33A to these cell types was also displaced by both unlabeled N33A and rhSDF-1␣ (not shown). Consistent and considerable migration of monocytes and T cells was induced by N33A (Fig. 2). The potency and efficacy of N33A in the induction of mononuclear cell migration was comparable with rhSDF-1␣, suggesting that chemically synthesized N33A retains the tertiary structure and functions as well as rhSDF-1␣.
The effect of SDF-1 on neutrophils is controversial (1, 3, 4). While some investigators failed to detect chemotactic activity of SDF-1 for neutrophils (1), others were able to induce significant Ca 2ϩ mobilization in neutrophils at physiologically relevant concentrations of SDF-1 (4). In an effort to clarify the activity of SDF-1 on neutrophils, we tested the binding and function of N33A on neutrophils in comparison rhSDF-1␣. As shown in Fig. 3, human peripheral blood neutrophils expressed a substantial number of specific binding sites for both N33A (Fig.  3A) and native SDF-1␣ (Fig. 3B). The binding is of high affinity with estimated K d values of about 5 nM. N33A and rhSDF-1␣ competed equally well for each other's binding as shown by the displacement curve ( Fig. 3C and data not shown). Neutrophils also migrated in response to both N33A and rhSDF-1␣, indicating that neutrophils are indeed among the target cell types for SDF-1.
To further confirm that N33A utilizes CXCR4 as its functional receptor, HEK293 cells stably expressing CXCR4 (CXCR4/293 cells) were employed. Wild type HEK 293 cells exhibited a low level of specific binding (about 200 binding sites/cell) for both N33A and native SDF-1␣. Both N33A and native SDF-1 also induced a weak but significant directional migration of wild type HEK293 cells (CI ϭ 2.1 Ϯ 0.2, at 120 nM ligand concentration). This is in agreement with the notion that a great variety of nonhematopoietic cells express CXCR4 mRNA. However, HEK293 cells overexpressing CXCR4 expressed markedly increased number of specific binding sites for radiolabeled N33A and rhSDF-1␣ (Fig. 4, A and B). N33A and rhSDF-1␣ mutually competed for binding to CXCR4/293 cells (Fig. 4, C and D). Both N33A and rhSDF-1␣ were able to induce a remarkable directional migration of CXCR4/293 cells (Fig. 5) with similar potency and efficacy. HEK293 cells transfected with known chemokine receptors, including CXCR1, CXCR2, and CCR1-5, did not show any increased binding or activation by N33A or rhSDF-1␣ over background levels, whereas these cells specifically bound and migrated in response to their ligands (data not shown). These results demonstrate that N33A, like rhSDF-1␣, uses CXCR4 as a functional receptor.
Chemokines are important mediators that participate in a variety of pathophysiological conditions including inflammation, infection, tissue injury and repair, immune responses, as well as malignancy (6,7). Recently members of chemokine receptor family have been identified to be fusion co-factors for HIV-1 entry into host cells. Some chemokines were able to inhibit HIV-1 entry through competitive occupancy of the relevant receptors such as CCR5 or CXCR4/Fusin/LESTR which mediate the fusion of either monocytotropic or T lymphocytotropic viruses, respectively, with CD4 ϩ host cells (2,3,13,(23)(24)(25)(26)(27). SDF-1 has recently been identified as the ligand for CXCR4 and was able to inhibit the cell fusion mediated by the envelope protein of the T lymphotropic HIV-1 strain (23)(24)(25)(26)(27)(28).
SDF-1 is a member of the CXC chemokine subfamily and was first identified as a molecule possessing pre-B cell-stimulatory activity (29,30). It was later described as a chemoattractant for resting T lymphocytes and monocytes (1). As a CXC chemokine, SDF-1 has several unique features in comparison to other  2. Chemotactic response of monocytes and T cells to N33A and rhSDF-1␣. Different concentrations of N33A or rhSDF-1␣ were placed in the lower wells of the microchemotaxis chamber. The cells were placed in the upper wells, which were separated from the lower wells by polycarbonate filters. After incubation the filters were removed, stained, and the cells that migrated across the filters were counted. The results are expressed as the CI representing the fold increase of cell migration in response to chemokines versus medium control. CI Ն 2 are statistically significant in comparison with the spontaneous migration (in response to control medium alone, CI ϭ 1). members of the same family. SDF-1 mRNA is expressed constitutively in virtually every tissue including heart, liver, lung, brain, muscle, spleen, and kidney (28 -30). Expression of SDF-1 gene is not affected by proinflammatory stimuli (31), in contrast to most other chemokines which are mainly expressed in response to proinflammatory cytokines and are believed to regulate the recruitment and activation of mature leukocytes at inflammatory foci (6,7). In the absence of inflammation, blood monocytes constantly replace mononuclear phagocytes in the tissue, sustaining a stable level by extravasation from blood stream and undergoing differentiation into macrophages. SDF-1 displays a tissue distribution that is considered appropriate for function in lymphocyte recirculation, in basal recruitment of monocytes and in normal replenishment and turnover of tissue mononuclear phagocytes (1, 30 -32). However, neutro-phils do not normally infiltrate organs or tissues even though these cells express binding sites for and migrate in response to SDF-1␣ in vitro, as further demonstrated in the current study. Thus, the role for SDF-1 as a primordial chemokine regulating primarily the tissue distribution of leukocytes needs further investigation.
SDF-1 also has several essential functions in development (2). Mice lacking SDF-1 due to homologous recombination died perinatally and although the numbers of B cell progenitors in mutant embryos were severely reduced in fetal liver and bone marrow, myeloid progenitors were reduced only in the bone marrow not in the fetal liver, indicating that SDF-1 is responsible for B cell lymphopoiesis and bone marrow myelopoiesis. In addition, mice deprived of SDF-1 gene had a cardiac ventricular septal defect (2).
The SDF-1␣ analogue N33A was chemically synthesized and was shown in this study to be equally as potent as native SDF-1␣ in binding and activating human leukocytes as well as CXCR4-transfected HEK293 cells. N33A has also been shown to induce human B lymphocyte migration and to effectively inhibit infection of PM-1 cells by HIV-1(LAV). 2 The preparation of analogues is greatly facilitated by molecular chemical synthesis, in which proteins can be produced either singly or by combinational methods. The analogues could include a full range of genetically encoded amino acids as well as unnatural backbone structures and unclonable residues such as D-amino acids, fluorescent or nuclear magnetic resonance-sensitive nuclei. The activities of analogues can also be tuned by fast cycles of synthesis-design-assay-resynthesis (33). The validity of this approach is demonstrated by the fact that it yields functional molecules such as the SDF-1␣ analogue N33A, a biological . CXCR4/293 cell migration was measured by 300 min incubation at 37°C in microchemotaxis chambers as described previously (21). CI Ն 2 are statistically significant in comparison with the spontaneous migration.
contamination-free agonist of SDF-1 that can be utilized in studies of the immune system and host defense against AIDS.