Skip Navigation


Carcinogenesis Advance Access originally published online on October 18, 2005
Carcinogenesis 2006 27(4):693-707; doi:10.1093/carcin/bgi235
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
27/4/693    most recent
bgi235v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Simmons, D. P.
Right arrow Articles by De Luca, L. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Simmons, D. P.
Right arrow Articles by De Luca, L. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Evidence that sequence homologous region in LRAT-like proteins possesses anti-proliferative activity and DNA binding properties: translational implications and mechanism of action

Denise Perry Simmons 1, Megan L. Peach 2, Jonathan R. Friedman 1, Michael M.B. Green 1, Marc C. Nicklaus 2 and Luigi M. De Luca 1, *

Laboratory of Cellular Carcinogenesis and Tumor Promotion, Center for Cancer Research, National Cancer Institute, NIH, Bethesda, MD 20892, USA, 1 Basic Research Program, SAIC-Frederick, and 2 Laboratory of Medicinal Chemistry, Center for Cancer Research, NCI-Frederick, Frederick, MD 21702, USA

* To whom correspondence should be addressed. Tel: +1 301 496 2698; Fax: +1 301 496 8709; Email: delucal{at}mail.nih.gov


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 DISCUSSION
 References
 
LRAT (lecithin:retinol acyltransferase), an enzyme whose levels are modulated during malignant conversion, has been reported as the founder member of a new LRAT-like family that includes tumor suppressors TIG-31–164 and Ha-Rev1071-162. The mechanisms that link these three proteins to carcinogenesis as well as the significance of a reported shared sequence homologous region remain unclear. This begs the question if the tumor suppressors possess enzyme properties and/or if the LRAT enzyme possesses tumor suppressor properties. We use the reported homologous region as a first approach to address the question from the perspective that all three proteins can possess tumor suppressor properties. We postulated that the homologous sequence harbors an anti-proliferation domain within the full-length proteins and that dodecapeptides of this sequence possess anti-proliferative activity. We report that H-TIG-3111–123, H-Ha-Rev107-1111–123 and H-LRAT160–171:C168L exhibited in vitro growth inhibitory activity in a human cutaneous melanoma (HCM) model and affected tumor growth in a nude mouse model. Further, in peptide-sensitive HCM cells, these peptides crossed the plasma membrane and localized to the nucleus, where they could bind and activate promoters of transcription factors involved in G1->S transition. Moreover, peptide-induced abrogation of cyclin dependent kinase-2 expression was concomitant with sub-cellular re-distribution of cyclins E and A. Indeed, the sequence homologous region within each full-length wild-type protein as well as the growth inhibitory peptides can form alpha helices, a likely configuration for binding to DNA. This is the first report that this sequence homologous region (AA111–123) within these LRAT-like proteins harbors an anti-proliferative domain with DNA binding properties. Sequences from this sequence homologous region can be used as templates for anti-tumor drug design and as probes to investigate disease-related mechanisms and structure-activity relationships of the full-length proteins, TIG-31–164, Ha-Rev1071–162 and LRAT160–171.

Abbreviations: cdk2, cyclin dependent kinase-2; EMSA, Electrphoretic mobility shift assay; FACS, fluorescence activated cell sorter GIP, growth inhibitory peptide; HMG A1, high mobility group protein; LMW, lower molecular weight; LRAT, Lecithin:retinol acyltransferase; RAR, retinoic acid receptor; RAREs, RAR elements; TGFß-1, transforming growth factor beta-1


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 DISCUSSION
 References
 
Tumor suppressors and transferases have been linked to carcinogenesis, and studies that investigate the relevance of this link have led to elucidating mechanisms of the disease process and to facilitating development of treatments. In fact, dual-functionality of proteins as relates to cell growth and carcinogenesis has been demonstrated for several enzymes; for example, P-TEN, an enzyme with phosphatase activity and glycosyltransferases, EXT1 and EXT2, have been shown to possess both enzyme and tumor suppressor properties (1,2). Furthermore, design and synthesis of peptides from their full-length proteins to yield therapeutic agents has proven successful. For example, peptides derived from laminin displayed cell growth modulating activity (3). Rational designed peptides have been shown to interrupt de-regulated Ras signaling (4); synthetic peptides have been shown to selectively target cancer cells (5) and to be used for vaccine development (6).

Lecithin:retinol acyltransferase1–230 (LRAT), classified as an enzyme that esterifies retinol through transfer of an acyl group from lecithin, shares no sequence homology with other known acyltransferases (7). However, LRAT recently has been shown to have its phylogenetic origins in a diverse family of proteins (8), and it is this diversity that underlies our question of possible LRAT function in addition to its enzymatic activity. In fact in the NIpC/P60 superfamily of enzymes, LRAT-like family members are classified by a circular permutation and a conserved cysteine (8). In this regard, among the members of the LRAT family are the two Class II tumor suppressors, human tazarotene induced gene-3 (H-TIG-31–164) and human Harvey ras revertant 107-1(H-Ha-Rev107-11–162) (9,10). These tumor suppressors have a 27 and 25% sequence identity to LRAT, respectively, and all family members share a region 12 amino acids long of very high sequence homology (Table I). Further, LRAT, Ha-Rev107 and H-TIG3 genes are retinoid responsive, thus linking these proteins to the retinoid specific receptors, RAR-{alpha}, -ß, -{gamma}, through which the retinoids mediate their effects (11). Relevant to our study those studies that demonstrated reduced LRAT enzyme activity and/or expression associated with different types of cancer, such as prostate, renal and breast (12–14). In human skin, the LRAT enzyme is functional in the basal cell layer where it catalyzes retinyl ester synthesis (15). Studies from our lab indicate that, in vitro, normal human proliferating melanocytes express the LRAT protein and are capable of retinol esterification by this enzyme (16). In addition, the Class II tumor suppressors, H-Ha-Rev107 and TIG, are frequently downregulated in tumorigenic cell lines and psoriasis, basal carcinomas and some aggressive squamous cell carcinomas (17). This protein/gene-associated aberrant activity of the LRAT enzyme and these two tumor suppressors in carcinogenesis, the reports in the literature of the dual functionality of enzymes, the fact that these three proteins are retinoid responsive and the recent report of these three proteins as members of the same family of proteins, combined with the reports of the shared 12-amino acid sequence homology of these tumor suppressors with LRAT (7,9), prompted us to ask if this sequence might be an anti-proliferation domain and if wild-type dodecapeptides synthesized from this homologous region would possess growth-inhibitory properties. If so, the sequence homologous region could provide a clue to the relevance of the LRAT-associated aberrant behavior in cancer.


View this table:
[in this window]
[in a new window]
  		Table I. Amino acid sequences of wild-type full-length proteins H-LRAT1–230 (AF 071510), H-Rev107-11-162 (X92814 [GenBank] ) and H-TIG-31–164 (AF 060228)

 
Cell cycle, particularly control of G0/G1->S, is intimately associated with carcinogenesis and the retinoids and their receptors, through modulation of retinoid responsive genes, are critical regulators of cell growth and differentiation. For instance, retinoic acid receptor-beta (RAR-ß), which is downregulated in some cancers, has been shown to modulate expression of the retinoid responsive cyclin E (18). The G1 cyclins, E and A, have as their cognate kinase, cyclin-dependent kinase 2 (cdk2). Expression levels as well as activity of these complexes are known to be responsible for checkpoint control, the so-called R-point, as the cell transits through to the S-phase. Indeed, it is the R-point that is said to play a critical role in inhibiting malignant transformation (19). Numerous transcription factors interact in coupling the G1/S machinery to the initiation of DNA replication. However, the transcriptional link between the cyclins and DNA synthesis/replication is the E2F family of transcription factors. This family is necessary for the formation of pre-replication complexes, requires interaction with cdk2 products, and interestingly has been shown to have a role in G1/S progression similar to that of Myc (20).

Human melanoma is one of the models of choice to investigate tumor progression toward metastasis (21). Use of this model also permits one to investigate approaches that may be clinically useful in cutaneous melanoma, which accounts for >77% of all deaths from skin cancer in the USA (22) and is the fastest rising cancer in developing countries (23). Studies from our lab have shown that melanocyte transformation results in two morphological phenotypes, which express LRAT protein but possess divergent esterification properties. Thus, in sharp contrast to other LRAT-associated cancers, epithelioid melanoma cells esterify retinol but fibroblastoid melanoma cells do not (16). In a continued effort to elucidate the significance of the associated aberrant expression/activity of LRAT in carcinogenesis, we used a human cutaneous melanoma model to ask whether the sequence homologous region found in the wild-type full-length proteins H-LRAT160–171, H-TIG-31–164, and H-Ha-Rev107-11–162 functions as an anti-proliferative domain, and if dodecapeptides of these sequences can function as growth inhibitors.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 DISCUSSION
 References
 
Peptide synthesis and fluorescein modification
All peptides were synthesized by SIGMA-GENOSYS (Woodlands, Texas). On receipt, lyophilized peptides were solubilized in sterile ultra pure water (KD Medical, Columbia, MD) to achieve a 10 mM stock, aliquoted and stored at –80°C or immediately used. Peptides H-LRAT160–171, H-Ha-Rev107-1112–123 and H-TIG-3112–123 were designed based on the published 12-amino acid sequence homology of their respective full-length wild-type proteins (9). An additional N-terminus amino acid, from the published full-length sequences of H-Ha-Rev107-11–162 or H-TIG-31–164, was added to the N-terminal sequences to facilitate purification. Thus, amino acid 111 E or R was added, respectively (Table I). Placement of the fluorescein-label and fluorescein labeling of peptides were determined based both on structural bioinformatics and fluorescein label chemistry (24). Flc denotes the fluorescein label placed at the N-terminus of the peptide (flc-NCEHFVTYLRYG = fluorescein labeled H-LRAT160–171:C168L; flc-RNCEHFVAQLRYG = fluorescein labeled H-TIG-3111–123). Unrelated peptide (GYFHEGFHGYFGY) was designed based on amino acids in the dodecapeptide that were not part of the tri-peptide sequence (AQL, AQC, NEL, TYL) distinct to each peptide. Active H-LRAT (H-LRAT160–171:C168L) and inactive H-TIG3 (H-TIG3111–123:L120C) peptides were designed based on the rationale that follows.

Rational design H-LRAT160–171:C168L
LRAT enzyme and the two tumor suppressor proteins share their highest degree of sequence homology within a 12-amino acid region/dodecapeptide. Within this region are four features on which we focused: a tri-peptide sequence unique to each 12-amino acid region, a leucine residue common to the tumor suppressors within their unique tri-peptide sequences, a cysteine residue in the LRAT dodecapeptide at the same position as the leucine residue in the tumor suppressor dodecapeptide and a C-terminal tri-peptide that is juxtaposed to the unique tri-peptides and is common to each 12-amino acid region (Table I). We predicted that (i) the 12-amino acid sequence of the full-length tumor suppressors harbors a domain that lends tumor suppressor properties to these peptides, (ii) leucine is important to the postulated anti-proliferative activity because both dodecapeptides of the full-length tumor suppressors contain a leucine (L120) at Position 9, which is within the tri-peptide sequence distinct to each peptide, (iii) H-LRAT160–171 would be inactive as an anti-tumor agent because at Position 9 wild-type H-LRAT160–171 contains a cysteine (C168), one of the two cysteines shown to be important although not essential to optimal enzyme activity of the full-length protein (7), while a cysteine to leucine (C168L) mutation would yield an active HLRAT:C168L growth-inhibitory agent and that (iv) if the former were true then a mutation in L120 (L120C) of H-TIG-3111–123 or H-Ha-Rev107-1111–123 peptide would result in an inactive anti-proliferative peptide (Figure 1). Chemical structures were generated, checked and cleaned-up in ChemDraw 6.0 (CambridgeSoft Cambridge, MA).


Figure 1
View larger version (19K):
[in this window]
[in a new window]
  		Fig. 1. Rational design of growth inhibitory H-LRAT160–171:C168L peptide. (A) Chemical structure of 12-amino acid wild-type peptides and predicted anti-proliferation activity (inactive versus active) in human cutaneous melanoma model (HCM). Sequence number is taken from the full-length wild-type proteins (Table I). N-terminal amino acids in bold italics added to facilitate synthesis. Underscored sequences are the three amino acids that distinguish each of the three peptides. Highlighted chemistry reflects the underscored three-amino acid sequences. Arrows indicate the position of the residues that are predicted to be important in anti-proliferation activity (L120) or known to be important in LRAT1–230 enzymatic activity (C168). (B) Predicted growth inhibitory activity and chemical structures of wild-type H-LRAT160–171 peptide and of cysteine to leucine mutation (C168L) in wild-type H-LRAT160–171 peptide. C168 in wild-type LRAT1-230 is in same position as L120 in wild-type H-Rev107-11-162 and H-TIG-31–164. Arrows point to site of mutation. (C) Chemical structure of leucine to cysteine mutation (L120C) in wild-type H-TIG3111–123 peptide and predicted growth inhibitory activity. Arrows indicate site of mutation.

 
Cell culture
Melanoma cell lines [American Type Culture Collection (ATCC)] were maintained as specified by ATCC (Rockville, Maryland). These are E:Hs939.T at passage 9, a malignant melanoma cell line, and the matched tumor pairs [F:688(A).T at passage 8, F:688(B).T at passage 5], [E:WM-115 at passage 73, E:WM-266-4 at passage 82]. Each pair was obtained from a single patient and consists of cells established from the primary site tumor and cells from a tumor that metastasized from the primary site. Healthy human primary epidermal melanocytes (HFSC/2), passage 2, obtained from Yale Skin Diseases Research Center (New Haven, CT), were maintained in melanocyte growth medium from Clonetics (San Diego, CA). All cell types were adult tissue in origin. E and F designations are melanoma with epithelioid or fibroblastoid morphology, respectively (16).

Assessment of cell proliferation and cell death
Cultures were set up in parallel for cell counts, western, and fluorescence activated cell sorter (FACS) analyses and were seeded, in duplicate or triplicate, in 6-well dishes to achieve 60–70% confluence overnight. The next day, monolayers were either mock-treated (water as vehicle), untreated, or treated with 1 µM peptide final/day, then harvested at 0, 24, 48, 72 and 96 h. The 96-h cultures were not included in the datasets because vehicle paired cultures consistently reached confluence between 72 and 96 h, and by 96 h there was light microscopy evidence of cell death in these vehicle and untreated cultures. Monolayers from each well and its respective medium were counted for live and dead cells using a hemacytometer and trypan blue staining. Monolayers were also analyzed by FACS. Each experiment was performed a minimum of three times.

Human tumor xenograft mouse model
E:WM-115 sub-confluent monolayers were harvested and resuspended in phosphate-buffered saline (PBS) just before use. A total of 20, six-week-old, Balb/c nu/nu female mice were weighed and inoculated subcutaneously in the right flank with 0.1 ml PBS or 5 x 106 cells in 0.1 ml PBS. Mice were grouped as 5 mice for mock-treated, 15 mice for growth inhibitory peptide (GIP) treatment which was subdivided into 5 mice/GIP. Naive animals were maintained to check for spontaneous tumor formation. Tumor volume was determined according to the formula: [(w)2 (l)]/2 (25), and tumor count was based on the observed development of one tumor/mouse. Tumors were allowed to reach a size that would permit evaluation of changes in tumor size. To assess peptide efficacy as an anti-tumor agent, a 0.05 ml GIP solution was administered as a single treatment, three-site intra-tumoral injection of 0.05 mg/kg body wt. Three single treatments were given over a 10-day time period. Animals were weighed weekly and tumor size measured by calipers on the day of treatment (Day 1) and every other day after treatment. At the end of the treatment protocol, two animals with excessive size tumors were treated daily, over a 4-day period, with H-Ha-Rev107-1111–123.

Western analysis
At each specific time point, peptide-treated and vehicle-treated monolayers were lysed with SDS lysis buffer. 20 µg of protein were loaded on a 4–12% Bis–Tris SDS–PAGE (Novex, Invitrogen, Carlsbad, CA) reducing gel, separated by electrophoresis, and transferred onto a 0.2 (m polyvinyldiene difluoride membrane. Blots were probed with polyclonal rabbit anti-actin (Chemicon International, Temecula, CA) and antibodies to HMGI(Y)/HMG A1, cyclin A, cyclin E, cdk2 and RARß1/2 (Santa Cruz Biotechnology, Santa Cruz, CA). Antibody specificity was verified using Santa Cruz control cell extracts and molecular weight standards. Specific signal was quantified by ImageQuant Analysis (Molecular Dynamics, Sunnyvale, CA). All experiments were performed in triplicate for a minimum of two times.

Subcellular localization
Subcellular localization of peptides was assessed using fluorescein labeled H-TIG-3111–123 (flc-RNCEHFVAQLRYG) or H-LRAT160–171:C168L (flc-NCEHFVTYLRYG) and real time confocal laser scanning fluorescence microscopy. Peptide localization was studied at peptide concentrations used in cell proliferation methods. Briefly, the most GIP-sensitive melanoma cells (E:WM115), the least GIP-sensitive melanoma cells (F:Hs688(B)T and the GIP-insensitive normal melanocytes (HFSC/2) were seeded, in triplicate, in 0.17 mm delta T dishes (Bioptechs, Butler, Pennsylvania ) and grown to 60–70% confluence. The cultures were placed on the 37°C temperature controlled stage, treated with vehicle, flc-H-TIG-3111–123 or flc-H-LRAT160–171:C168L, and images were continuously captured over 48 h. Additional images were captured at 72 and 96 h. During the last 30 min of the peptide treatment time-course, peptide-treated cells were incubated with the nuclear staining agent, Hoechst 33342 (Molecular Probes, Eugene, OR). Confocal fluorescent images were collected with a Bio-Rad MRC 1024 confocal scan head mounted on a Nikon Optiphot microscope with a 60x planapochromat lens then saved using Bio-Rad LaserSharp software. Experiments were performed in excess of three times. Line graph intensity profiles of fluorescence intensities were measured after acquisition with the Zeiss AIM software.

Transcription factor analysis
Transcription factor gene array analysis (Panomics Transcription factor pathfinder array TranSignal kit, Redwood City, CA) was used to profile activities of 54 transcription factors involved in signal transduction. Briefly, nuclear extracts were prepared using the Pierce nuclear extraction kit (Rockford, IL) from 72-h peptide and vehicle treated cultures and processed according to the manufacturer's instructions. Differential hybridization signals, corresponding to differentially bound transcription factors by the nuclear extracts, were captured by West Pico Chemiluminescence Imaging (Pierce Rockford, IL) and quantified in ImageJ. Specificity of GIPs binding was determined by titrating the peptide (32 ng/blot and 32 pg/blot) against the transcription factor labeled blots and using the optimal GIP concentration (32 ng) for a final specificity test of GIP (32 ng) versus scrambled peptide (32 ng) on each blot. Optimal GIP concentration was based on signal to noise ratios at each dilution. Values >1.5-fold above background were considered positive and ‘boxed’ as a binding signal. Experiments were performed a minimum of two times.

Electrphoretic mobility shift assay
Panomics electrphoretic mobility shift assay (EMSA) kit (Redwood City, CA) was used to verify select GIP-bound transcription factors. Briefly, vehicle-treated or GIP-treated nuclear extracts were incubated with the specific double-stranded biotinylated probe, that is DNA response elements [RXRE(DR1): AGGTCACAGGTCACAGGTCACAGGTCACAGGTCA or RARE(DR5):AGGTCACCAGGAGGTCACCAGGAGGTCACCAGGAGGTCACCAGGAGGTCA]. Each sample was separated on 6% native polyacrylamide gels and shifted bands that corresponded to protein/DNA complexes were captured by an HRP-based detection system. Experiments were performed in excess of three times.

Transcription factor cell reporter assay
The TranSignal Transcription Reporter Array kit (Panomics, Redwood, CA) was used to assess GIP-induced transcription factor activation of select GIP-bound transcription factors [EGR, E2F, NFAT, RARE (DR5), RXRE (DR1), TR (DR4), VDR (DR3)]. Briefly, 80–90% confluent E:WM-115 cells were transfected with the plasmid DNA mix then treated with vehicle (water), H-TIG-3111–123 or H- LRAT160–171:C168L. Specific signal was determined using a horseradish peroxidase chemiluminescence detection system and quantified using IMAGEJ. Arbitrary galactosidase units were determined by normalization against empty vector. Data are presented as the mean of two independent experiments with four replicates/construct/experiment.

Molecular models
The full-length sequences of the H-LRAT1–230, H-TIG-31–164 and H-Ha-Rev1071–162 proteins were submitted to 12 computer programs for protein secondary structure prediction. The programs used and their references are listed in Table II. In each case we submitted the individual sequence of each protein; however, many of the methods include an automatic sequence search to find and align homologous sequences to the query before making the prediction. Helical models for each peptide were generated using the Maestro interface for MacroModel (Schrödinger) and subjected to a small 500-step conjugate gradient minimization using the OPLS-AA forcefield (27) with continuum solvent (28).


View this table:
[in this window]
[in a new window]
  		Table II. Secondary structure predictions methods

 
Transient transfection of RAR-ß
PCMX-hRAR-ß2 was a generous gift of Ronald M. Evans, Gene Expression Lab (San Diego, CA). Plasmid was grown and purified according to Qiagen Endofree Plasmid protocol [Briefly, E:WM-115 cells were transfected with RAR-ß2-plasmid using Super-Fect (Qiagen, Valencia, CA)] Super-Fect alone, vehicle for 4 h at 37°C in 2% serum containing medium. Transfection solutions were removed and cells were maintained 48 h in 10% medium with and without 10–6 M 9-cis retinoic acid. Cells were viewed daily for morphological changes, harvested and processed as described in methods for cell proliferation. After conditions were established experiments were performed a minimum of two times.

Statistical analysis
The t-Test for the significance of the difference between the means of two independent samples was used to evaluate the effect of GIPs on normal human epidermal melanocytes. All P-values were two-sided and were considered statistically significant if P < 0.05. Standard deviations were calculated using Excel software program.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
DISCUSSION
 References
 
In vitro proliferation studies: proof of hypothesis
H-TIG-3111–123 1, H-Ha-Rev107-1111–123 2, and LRAT160–171:C168L 4 peptides inhibited proliferation of human melanoma cells, E:Hs939.T (not shown) and E:WM-115, [mean(%) ± SD] 60.9 ± 8.9, 42 ± 5.5 and 39.2% ± 2.8, respectively when used at 1 µM/day, over 48 h (Figure 2A) but the homologous wild-type H-LRAT160–171 dodecapeptide 3 did not (4.6% ± 1.7). Thus, the C168L mutation in the LRAT dodecapeptide resulted in a 9-fold increase in growth inhibition relative to the wild-type H-LRAT160–171 3. In sharp contrast, the L120C mutation in H-TIG-3111–123 resulted in an inactive peptide (2% ± 1.4) (Figure 2A), so that the growth inhibitory activity exhibited by the wild-type H-TIG-3111–123 peptide was decreased 30-fold, thus validating the significance of the leucine residue in growth inhibitory activity.


Figure 2
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 2. Wild-type peptide sequences from tumor suppressors and mutated LRAT peptide result in abatement of proliferation activity in human melanoma cells. (A) Y-axis: percent decrease in cell number affected by incubation of E:WM-115 (shown) melanoma cells with various peptides (X-axis) at 1 µM final concentration/day over 48 h. Unrelated peptide (GYFHEGFHGYFGY), H-Ha-Rev107-1111–123 2, H-TIG-3111–123 1, H-TIG-3111–123:L120C 5, H-LRAT160–171 3, H-LRAT160–171:C168L 4. Changes in cell number are depicted as percent. Actual cell number (B) Effect of H-TIG-3111–123 1 and H-Ha-Rev107-1111–123 2 on cell proliferation in normal proliferating human melanocytes at 24 and 48 h using same conditions as in (A). Bold printed numbers adjacent to peptide names refer to chemical structures in Figure 1.

 
Importantly, we observed that the anti-proliferative peptides were inactive in normal melanocytes. Evaluation of 24- and 48-h peptide-treated versus untreated HFSC/2 normal melanocytes did not show a significant change in doubling time of the cells (P < 0.05) (Figure 2B). Neither trypan blue staining nor Annexin V/propidium iodide FACS analyses showed evidence of cell death or changes in cell morphology in normal melanocytes >96 h or these anti-proliferative peptide treated melanoma cells within 72 h (data not shown).

GIP inhibit proliferation in cells lines derived from different stages of melanomagenesis
Proteins can be classified as tumor suppressors based on different properties, including structure, sequence, biological activity and function. The observed peptide growth inhibitory activity of H-Ha-Rev107-1111–123, H-TIG-3111–123 and H-LRAT160–171:C168L satisfies a criterion for tumor suppressor activity. Because H-TIG-3111–123 was the most effective of the three GIPs, it was used as the prototype to evaluate the ability of the GIPs to inhibit cell proliferation in cell lines representative of progression in melanomagenesis. 1 µM /day, over 48 h, of H-TIG3111–123 significantly inhibited proliferation of cells from the primary site tumor cell lines established from two different patients (E:WM-115 = 60% and F:Hs.688(A).T = 55%). The observed levels of inhibition of cell proliferation (Figures 2A and 3,) without evidence of apoptosis are equivalent to those reported in the literature for 0.1 ng/ml of transforming growth factor beta-1 (TGFß-1) in neuroendocrine tumor cells (29). Even though the H-TIG-3111–123 was less effective in cells from the metastatic matches, cells from the metastasized epithelioid tumor, E:WM-266-4 at 33%, were more sensitive to H-TIG-3111–123 than those of the metastasized fibroblastoid tumor, F:Hs.688(B).T at 8.3% (Figure 3). Hence, the epithelioid melanoma cell line is more responsive to GIP than the fibroblastoid phenotype; furthermore, cell lines from primary sites tumors are more sensitive to GIP than their metastatic matches, regardless of morphological phenotype.


Figure 3
View larger version (20K):
[in this window]
[in a new window]
  		Fig. 3 Effect of H-TIG-3111–123 1, 1 µM final/day over 48 h, on proliferation of cells from matched tumors of primary and metastatic origin. Y-axis: Percent decrease in cell number, X-axis: cell lines established from melanomas: patient 1, E:WM-115 an epithelioid primary site tumor and E:WM-266-4 metastatic match; patient 2, F:Hs.688(A). T fibroblastoid primary site tumor and metastatic match F:Hs.688(B).T.

 
Human xenograft nude mouse model indicates that GIPs affect in vivo growth
The in vitro anti-proliferation activity of the GIPs prompted us to ask whether these peptides inhibit tumor growth. Subcutaneous human xenografts of the peptide sensitive E:WM-115 cells in Balb/c nu/nu female mice were used to assess in vivo growth inhibitory properties of the GIPs. One palpable tumor mass/mouse (n = 16) appeared at the grafting site within 1–3 months (Table III). Of the 20 mice that received subcutaneous melanoma cell grafts, 3 mice did not develop tumors, 1 of these died post grafting, and 1 mouse had a palpable tumor at the end of the study. At the end of the treatment protocol, change in the size of each tumor was determined based on its measured size on Day 1 of treatment relative to its measured size on Day 12 following the last treatment. Vehicle treated tumors (n = 5) increased in size an average of 17.8-fold with a range of +3.4–+44.7. The +3.4-fold increase in tumor size in the vehicle control group was used as the lower limit tumor size cut-off to assess GIP efficacy. Thus, GIP-treated tumors that increased in size ≥+3.4 at the end of the protocol were considered non-responsive to GIP. In the GIP-treated tumors (n = 11), responsive GIP-treated tumors (n = 9) were 3.1–4.5-fold smaller. In 5 of 11 tumors (nos 6, 7, 11, 12 and 16), the actual size of the tumor was reduced while tumor size was contained for 4 of 11 tumors (nos 8, 13, 17 and 18). Two of the GIP-treated tumors (nos 9 and 14) were resistant to GIP treatment. Growth inhibitory effects lasted 12 days after the final treatment, at which time tumor cell growth increased up to 4-fold (data not shown). To test the observed cytostatic effect (nos 8, 13 and 17) of GIPs, mice with the largest tumors (>500 mm3), one from the H-TIG-3111–123 treatment group (no. 9) and one from the PBS-treatment-group (no. 5), were treated with daily intra-tumoral injection over 4 days with 0.15 mg/kg body wt/0.05 ml/tumor H-Ha-Rev107-1111–123. As a result tumor growth was blocked, and this block was lifted within 7 days after the last intra-tumoral injection (data not shown). These in vivo data suggest that the GIPs possibly behave as anti-neoplastic agents; however, a larger study would need to be conducted to validate this. We next attempt to elucidate the mechanism of the action of the GIPs by investigating their (i) cellular and molecular interactions, (ii) intracellular localization and (iii) structure.


View this table:
[in this window]
[in a new window]
  		Table III. Effects of GIPs in in vivo human xenograft nude mouse model

 
GIPs' effect on G1->S cell cycle proteins
Because all three GIPs possessed growth inhibitory activity in melanoma cells, we studied the effect of the GIPs on candidate proteins common to both carcinogenesis and cell cycle molecular pathways. It is known that the retinoids, their receptors and specific retinoid responsive genes, such as LRAT, are part of a regulatory feed-back loop. Further, since the wild-type full-length protein H-TIG-31–164 is induced by the RAR-ß/{gamma} selective ligand, tazarotene, both H-TIG-31–164 and H-LRAT1–230 being retinoid responsive, and RAR-ß has been shown to be downregulated in some cancers, we hypothesized that RAR-ß would be a GIP molecular target in peptide sensitive tumor cells. RAR-ß basal expression levels were ~3-fold less in epithelioid melanoma cells, E:Hs939.T, relative to normal proliferating melanocytes (HFSC/2). Incubation with GIPs induced upregulation of RAR-ß protein in these melanoma cells to levels observed in normal proliferating melanocytes. A similar phenomenon was observed for the chromatin remodeling protein/transcription factor HMG A1, a high mobility group protein, specifically expressed in skin and frequently associated with neoplastic transformation/metastasis (Figure 4A). HMG A1 contains retinoic acid receptor elements (RAREs) in its promoter and is known to interact with transcription factors nuclear factor kappa-B (NF{kappa}-B), RAR and nuclear factor of activated T cells (NFAT)(30).


Figure 4
View larger version (24K):
[in this window]
[in a new window]
  		Fig. 4. Effect of GIP on expression of key cell cycle and transcription factor proteins. (A) RAR-ß and HMG A1 response in GIPs in E:Hs939T melanoma cells. Cells were exposed to 1 µM final/day over 48 h of GIPs, H-TIG-3111–123 1 or H-Ha-Rev 107-1111–123 2, then protein expression of transcription factors RAR-ß and HMG A1 were assessed in vehicle-treated normal melanocytes and vehicle- and GIP-treated melanoma cells. Transcription factor protein expression levels in arbitrary units, normalized against actin; as average of two separate experiments of replicate samples. Boxed: representative gel. Lane 1, E:Hs939T treated with H-HaRev107-3111–123 2; lane 2, E:Hs939T treated with H-TIG3111–123 1; lane 3, E:Hs939T with vehicle; lane 4, normal melanocyte with vehicle. (B) Western analysis of cell fractionated melanocytes, E:WM-115 vehicle-, H-TIG-3111–123- or H-LRAT160–171:C168-treated and analyzed for cdk2, cyclin-E and cyclin-A. (C) Drawing to depict summary of cellular distribution of cell cycle proteins based on Figure 4B. Outer shell, cytoplasm; inner shell, nucleus; A and E, cyclin A and cyclin E, respectively; E lmw, cyclin E lower molecular weight products; cdk2, cyclin dependent kinase 2. This drawing does not reflect the expression levels of these proteins.

 
We then assessed the expression levels and cellular distribution of cyclin E, a cell cycle regulator that is modulated by RAR-ß and, like RAR-ß, contains an RARE (DR5) and is important in G1->S (31,32). Moreover, cyclin E-cdk2 complexes found in the nucleus of normal cells are essential for timely entry into S-phase in some cell types (19). Cyclins E and A were found to partition in the nucleus of normal melanocytes, with cdk2 found in both nuclear and cytoplasmic fractions (Figure 4B). Melanoma cells, on the other hand, showed partitioning of cyclins E and A into both cytoplasmic and nuclear fractions with restriction of cdk2 to the cytoplasmic fraction. GIPs were effective in re-partitioning of cyclin A into the nucleus in melanoma cells, and incubation with H-LRAT160–171:C168L resulted in re-distribution of cyclin E to the nuclear compartment. Curiously, H-TIG-3111–123-treated melanoma cells showed both cytoplasmic and nuclear distribution and induced appearance of a lower molecular weight (LMW) form of the cyclin E in the nucleus. It has been reported that only tumor cells have the machinery, including the protease calpain, to process cyclin E into these LMW forms (33). Hence H-TIG-3111–123 possibly induces upregulation/expression of calpain.

The observed GIP-induced cyclin re-distribution can play a role in achieving a ‘normal’ cell cycle profile, but cyclin protein expression is also critical. After normalization against actin, we found that total cyclin E levels were elevated 15-fold in the vehicle-treated E:WM-115 melanoma cells relative to the normal melanocytes and this increased expression was reduced, albeit to a small extent by GIP treatment. Total cyclin A was over-expressed as well in the E:WM-115 melanoma cells relative to normal melanocytes, and this expression also was slightly downregulated by H-TIG-3111–123 and unaffected by H-LRAT160–171:C168L. Interestingly, in these same melanoma cells, total cdk2 expression was abrogated by LRAT160–171:C168L and significantly downregulated by H-TIG-3111–123 with detectable amounts in both nuclear and cytoplasmic fractions (Figure 4B). Figure 4C depicts a summary of the distribution of these important G1->S cell cycle proteins.

GIPs move to the nucleus in peptide sensitive melanoma cells
Modulation of transcription factors RAR-ß and HMGA1, of the G1 cyclins A and E, of cdk2, as well as the observed differences in sensitivity of the matched tumor pairs to the GIPs compelled us to investigate the subcellular localization of these GIPs. Fluorescein-labeled H-TIG-3111–123 or H-LRAT160–171:C168L combined with laser imaging confocal microscopy were used to assess the ability of the GIPs to cross the cell membranes of GIP-sensitive and insensitive cells. Sensitivity was based on proliferation data (Figure 3). Time course experiments with these fluorescein-labeled peptides showed that, within 48 h, at all concentrations, both H-LRAT160–171:C168L and H-TIG-3111–123 (Figure 5A) crossed the plasma and nuclear membranes and localized to the nucleus of the GIP-sensitive primary site epithelioid melanoma cells, E:WM-115. However, in the less sensitive fibroblastoid metastatic cells, F:Hs.688(B).T, H-LRAT160–171:C168L (Figure 5B.1 and B.2) or H-TIG-3111–123 (Figure 5B.3) localization was restricted to the cytoplasm. In sharp contrast, the normal proliferating melanocyte [field shown in both phase contrast (left panel) and fluorescence] fluorescein labeled GIP was not found in the cell (Figure 5C). Both the scrambled peptide and the inactive wild-type LRAT peptide were able to cross the plasma/nuclear membranes in peptide sensitive cells, but in the case of the former distribution was throughout the cell (data not shown). Line graph intensity profiling of the tumor cells captured in Figure 5A and C clearly demonstrates peptide nuclear localization in epithelioid tumor cells versus cytoplasmic localization in fibroblastoid tumor cells (Figure 5D). The observed cellular distributions were the same at each time point captured.


Figure 5
View larger version (44K):
[in this window]
[in a new window]
  		Fig. 5. Sub-cellular localization of fluorescein-labeled H-LRAT160–171:C168L 4 and fluorescein-labeled H-TIG3111–123 1. Sensitivity determination is based on proliferation data in Figure 3. (A) Peptide-sensitive E:WM-115 melanoma cells. Panels (1–3): Arrows point to in-sets of designated cells to illustrate imaging techniques. (1) Hoescht 33342 stained nuclei blue. (2) Fluorescein-labeled H-LRAT160–171:C168L 4 [green] was found sequestered to nucleus within 48 h. (3) Merge of Hoescht stained, fluorescein-labeled treated cells, that are shown in in-sets, demonstrated nuclear localization [blue] of fluorescein-labeled H-LRAT160–171:C168L 4. Panels (4–6) without using Hoescht 33342: H-TIG3111–123 1 time course of peptide entry and nuclear localization; in-set shows an improved resolution of the cell. (B) Peptide–insensitive F:Hs688(B).T melanoma cells. (Panel 1) Merge of Hoescht stained, fluorescein-labeled H-LRAT160–171:C168L 4 treated cells after 4 h incubation with peptide. (Panel 2) Fluorescein-labeled H-LRAT160–171:C168L, without using Hoescht 33342, was restricted to the cytoplasm during the 48-hour incubation. (Panel 3) (without using Hoescht 33342): H-TIG-3111–123 1 peptide entry into cells and restriction to cytoplasm during the 48-hour incubation. (C) Melanocytes (HFSC/2). (Panel 1) phase contrast to demonstrate presence of H-TIG-3111–123 1-treated cells in (Panel 2) confocal image that demonstrated peptide did not cross the melanocyte cell membrane. (D) Line graph intensity profiling of fluorescein-labeled peptide and Hoescht distribution in epithelioid cells from Figure 5A–b3 and fibroblastoid melanoma from Figure 5B–c1. Scanning axis X->Z. Blue, Hoescht; green, fluorescein–labeled peptide.

 
GIPs bind to and GIP-treated nuclear extracts interact with transcription factor cis-binding elements of proteins involved in G1->S
GIP localization to the nucleus (Figure 5) and modulation of transcription factors, RAR-ß and HMG A1 (Figure 4), led us to speculate that the GIPs may affect growth by interacting with transcription complexes, either by binding directly to DNA or by interacting with protein–DNA complexes. Indeed, these GIPs may well be involved in DNA binding since they are basic peptides and basic peptides have been shown to enhance protein–DNA interactions (34). We used a ‘pathfinder’ DNA transcription array and cell fractionation of H-TIG-3111–123-treated E:WM-115 melanoma cells to address these questions. Nuclear extracts from vehicle-treated E:WM115 cells bound AP2, MEF-1, OCT-1, TR(DR4), VDR(DR3) and E2F cis-binding elements. Nuclear extracts from E:WM-115 melanoma cells treated 72 h with H-TIG-3111–123 are capable of selective complex formation with 14 of 54 DNA cis-binding elements which include SMAD 3/4, Pax-5, a unique subset [EGR, GRE, NFAT, RAR(DR5), RXR(DR1), and TR] as well as causing a 2-fold increase, relative to vehicle, in binding of TR (DR4), VDR (DR 3) MEF-1 and AP2 (Figure 6A). In fact, in a cell-free assay used to assess direct peptide binding to DNA, GIPs directly bound AP2, MEF-1, OCT-1, Pax-5, SMAD 3/4 and a unique subset Myc-max (Figure 6B). The scrambled peptide control, used at the same concentration as the GIP (See Materials and methods), did not bind any of these elements. A concept map of the bound transcription factors from the different treatment groups revealed their common and unique subset of transcription factor cis-acting elements (Figure 6C).


Figure 6
View larger version (28K):
[in this window]
[in a new window]
  		Fig. 6. GIP specific interaction with select transcription factor cis-binding elements. (A) Gene array analysis of transcription factor cis-binding elements bound by nuclear extracts from 72-h vehicle- and H-TIG-3111–123 -treated (16 ng/day) E:WM-115. (B) Gene array analysis of transcription factor cis-binding elements directly bound by H-TIG-3111–123. 32 ng peptide/blot is compared to scrambled peptide control to show GIP binding is specific. (C) Concept map of array analyses in A and B demonstrating binding patterns from different treatment groups. Note unique subsets of cis-binding elements and overlapping cis-binding elements. Gray circle: nuclear extract vehicle-treated E:WM-115; Dashed circle: nuclear extract H-TIG3111–123 -treated E:WM-115; Black circle: 32 ng H-TIG3111–123. (D) Names and sequences of cis-binding elements in A, B and C with which GIP and/or nuclear extracts interact.

 
We performed an EMSA to confirm the nuclear extract array data. Because GIPs were capable of upregulating RAR-ß protein expression (Figure 4A), which requires heterodimers of RAR and RXR, we were especially interested in the interaction of nuclear extracts from H-TIG-3111–123-treated E:WM-115 cells with RARE(DR5) and RXRE(DR1). Three different specific migrating binding complexes were found as determined by 100x competitor probes. In nuclear extracts from both the vehicle-treated and GIP-treated E:WM-115 cells, RARE(DR5) biotinylated probes specifically bound the slowest migrating complex, S1. However, in H-LRAT160–171:C168L-treated nuclear extract substantially more S1 complex was found than in vehicle-treated. There are only marginal differences in amount of S2 complex found across the three treatment groups, whereas, the fastest migrating complex, S3, is found in H-TIG-3111–123-treated extracts (Figure 7A). RXRE(DR1) binding pattern shows that a specific, S1, is uniquely found in nuclear extract from H-LRAT160–171:C168L-treated E:WM-115. S3, is found in nuclear extracts from H-TIG-3111–123-treated E:WM-115, while an S2 is common to all three treatment groups. In fact, we believe this RARE EMSA data can serve as a paradigm for what we propose could be a GIP-recruitment of different proteins to the transcription complex. Indeed, when similar RAR-ß EMSA binding profiles were reported in the literature, it was suggested that the different migrating complexes for RARE beta were due to sequentially occupied RXR homo-dimers and RAR–RXR hetero-dimers (35). Moreover, RXR can heterodimerize with VDR and TR. Our array concept map, which revealed unique and overlapping subsets of GIP alone compared to nuclear extracts from GIP-treated and vehicle-treated tumor cells, corroborates this. The order of binding, the response elements bound, and subsequent transcription are most likely dependent upon the protein milieu in the microenvironment, reminiscent of co-activator/co-repressors recruitment and interaction with response elements to modulate transcription through chromatin remodeling (36,37). Additional transcription factor response elements (AP2, Myc, Oct-1 and SMAD 3/4) were tested by EMSA and shown to validate their respective binding array results (data not shown).


Figure 7
View larger version (25K):
[in this window]
[in a new window]
  		Fig. 7. Transcriptional targets of GIP-induced promoter activity. (A) Nuclear extracts GIP-treated E:WM115 interaction with RARE(DR5) Lane 1, free biotinylated RARE probe; lane 2, RARE biotinylated probe+H-TIG-3111–123 nuclear extract; lane 3, RARE biotinylated probe+ 100x cold RAR probe+ H-TIG-3111–123 nuclear extract; lane 4, RAR biotinylated probe + vehicle nuclear extract; lane 5, RAR biotinylated probe+ 100x cold RAR probe+ vehicle nuclear extract; lane 6, RARE biotinylated probe +H- LRAT160–171:C168L nuclear extract; and lane 7, RARE biotinylated probe+ 100X cold RARE probe+ H-LRAT160–171:C168L nuclear extract. Specific RARE bound complexes (S1, S2, S3) (B) Nuclear extracts GIP-treated interaction with specific RXRE (DR1). Lane 1, free RXRE biotinylated probe; lane 2, RXRE biotinylated probe + H-TIG-3111–123 nuclear extract; lane 3, RXRE biotinylated probe+ 100X cold RXRE probe+ H-TIG-3111–123 nuclear extract; lane 4, RXRE probe + vehicle nuclear extract; lane 5, RXRE biotinylated probe + 100X cold RXRE probe+ vehicle nuclear extract; lane 6, RXRE+ H-LRAT160–171:C168L nuclear extract; and lane 7, RXRE biotinylated probe + 100X cold RXRE probe + H-LRAT160–171:C168L nuclear extract. Specific RXRE bound complexes (S1, S2, S3). (C) Nuclear extracts GIP-treated E:WM115 can activate transcription of bound transcription factor response elements. Black bars, H-LRAT160–171:C168L nuclear extracts; White bars, H-TIG-3111–123 nuclear extracts; each bar is the mean of two experiments with four replicates/construct/experiment. X-axis: arbitrary units galactosidase tagged reporter. Y-axis: cis-element reporter constructs.

 
GIPs activate transcription of transcription factors involved in G1/G0->S
We tested whether several of the GIP-induced transcriptional complexes observed in the array (Figure 6) are functional to drive target gene expression. GIPs drove transcription from the promoter region of Myc, E2F, RXRE(DR1) and RARE(DR5) (Figure 7C). As expected, the GIPs did not induce STAT3 reporter expression. However, VDR(DR3), which is bound by the GIP-treated nuclear extract, clearly is not inducible (Figure 7C). This may be a function of the cells' competition for RXR, which is the binding partner for vitamin D receptor 3 (VDR3), throid hormone receptor (TR) and RAR, the absence of the necessary co-activators to activate and/or the presence of co-repressors to inhibit VDR3 transcription.

GIPs are predicted to form alpha helices
We next addressed whether these peptides formed biologically relevant secondary structures. The three-dimensional structures of the full-length H-LRAT1-230, H-TIG-31–164 and H-Ha-Rev1071-162 proteins are difficult to predict, since they have no detectable homology to any other protein of known structure. As a first step toward understanding the structure-function relationship of the proteins and importantly their homologous dodecapeptide regions, we performed a series of computational secondary structure predictions on each sequence. These predictions indicated that the dodecapeptide region of each full-length protein sequence is a discrete secondary structural element because it is flanked on each side by regions of random coil (data not shown). Modeling predictions also suggested that the secondary structural element is likely to be an alpha helix (Figure 8A). This helix formation is of considerable interest since alpha helices are well-documented DNA binding structures (38) and it is this alpha helix configuration that we believe permits the GIPs to engage in the observed binding of DNA response elements. However, it is not clear from our models what significance the key leucine residue holds. We speculate that the helix residues potentially important in DNA contacts are within the terminal RYG (AA169–171/H-LRAT160–171:C168L and AA121–123/H-TIG-3 and H-Ha-Rev107-1) and the adjacent tri-peptide unique to each GIP[AQL118–120 (H-Ha-Rev), NEL118–120 (H-TIG3), TYL166–168 (H-LRAT)] (Figure 8B). This speculation is loosely based on a three amino acid sequence, NYL, found in zinc fingers in the region between the fingers, proximal to the D-box second zinc finger, within ‘M4’ (39). Interestingly, the single R residue in the ‘M5’ region, between the zinc fingers distal to the P-box of zinc finger 1, is shown to be a very important DNA contact in the nuclear receptor family. Furthermore, the C-terminal RYG amino acid sequence common to the GIPs is found in RXR and RAR DNA binding domains and has been shown by NMR, molecular modeling and crystallography to have points of contact with the DNA phosphates of RARE, such that 8 of 15 amino acids in RAR and 9 of 15 amino acids in RXR are R, Y, G, with a random spatial distribution (R-Xn-Y-Xn-G) (40).


Figure 8
View larger version (24K):
[in this window]
[in a new window]
  		Fig. 8. Molecular models of growth inhibitory peptides. (A) Helical models for each peptide were generated using the Maestro interface for MacroModel and subjected to a small 500-step conjugate gradient minimization using the OPLS-AA forcefield (27) with continuum solvent (28). Side View of the Key Leucine and the RYG (argine, tyrosine, glycine residues, respectively) possibly involved in DNA contacts. (B) Top view of GIPs, looking down the axis of the helix, highlighting the differences in sequence among the three GIPs. Note: the tri-peptide that contains each GIP unique dipeptide residues which are adjacent to the key leucine residue (AA118,119 and 120 for H-TIG3111–123 and Ha-Rev107111–123; AA166, 167 and 168 for LRAT160–171:C168L).

 

   DISCUSSION
Top
 Abstract
 Introduction
 Materials and methods
 Results
 DISCUSSION
References
 
LRAT has been reported as the founder member of a new class of enzymes (41) that includes proteins as diverse as tumor suppressors, including Ha-Rev107 and H-TIG-3, named LRAT-like proteins (8). We tested whether a sequence homologous region shared among these three proteins could have importance in understanding the mechanisms of their link to carcinogenesis and possibly implicate a shared function. We have used a human cutaneous melanoma model to demonstrate that the sequence homologous region can behave as a growth regulatory domain with DNA binding properties and that peptides of these sequences can be used to develop potential anti-tumor molecules as well as to investigate structure–activity relationships of full-length of H-LRAT, H-TIG-3 and H-Ha-Rev107. Most interesting is our observation that peptide sensitivity relates to morhphological phenotype and disease progression of the melanoma cell, that is, epithelioid versus fibroblastoid and primary tumor cell versus metastatic. Because we observed that one aspect of peptide-sensitivity in these two morphological phenoytypes was correlation with the ability of the peptide to cross the plasma and nuclear membrane, we believe it is worth investigating whether this difference in behavior of the peptides (or response of the cells to the peptides) in the two morphological phenotypes may reflect membrane differences in the two morphological phenotypes, for example, redox potential and/or the presence/absence of shuttle proteins.

Our results showed that in the primary site epitheiliod melanoma cells (E:WM-115) both cyclins E and A are overexpressed. We believe the observed cyclin E overexpression combined with its ectopic distribution of cyclin E can in part account for the hyper-proliferative nature of this tumor cell type (Figure 4C.2). This is reasonable since it has been shown that cytoplasmic cyclin E is capable of cdk2-independent S-phase entry (42) and that overexpression of cyclin E can result in accelerated S phase entry (43).

Since the GIPs were not capable of significant downregulation of the overexpressed cyclins E and A, the observed GIP-induced abrogation of their cognate kinase, cdk2 (Figure 4B), is not only advantageous but likely the key to the observed reduced proliferation (Figure 2A). Importantly, this level of regulation may have relevance because there is evidence that proliferation rates in human malignant melanoma is linked to clinical outcome (44), and the literature is replete with evidence for the role of cdk2 in delaying the S-phase.

Thus we suggest that the GIPs target multiple transcription factors involved in cell cycle regulation (Figure 9). The most significant consequence of this targeting is likely failed complex formation of cdk2 with cyclin E, which would result in a lower level of cdk2/cyclin E complex than is needed for triggering S-phase entry (19,43). Indeed, we found that GIP-treated primary site melanoma cells exhibited slowed growth without evidence of morphological differentiation or cell death by trypan blue staining and light microscopy (data not shown).


Figure 9
View larger version (18K):
[in this window]
[in a new window]
  		Fig. 9. Model of GIP-induced reduction in cell proliferation. Primary G1/S molecular targets believed to be responsible. Solid arrows: up-regulates expression; Dashed arrows: down-regulated expression, single-lined block arrow: inhibit complex formation; double-lined block arrow: inhibit G1 to S progression. double –lined rectangular boxes indicate predicted consequences of GIP-induced changes. GIPs can directly and indirectly affect transcriptional expression/activity of key regulatory target cdk2, which in turn impacts the replication/synthesis machinery through cyclin A and E2F and duration of the cell cycle through cyclin E.

 
Additionally, GIPs could affect an S phase arrest (Figure 9) in these peptide-sensitive melanoma cells. Failure of cyclin A–cdk2 complex formation would prevent late G1->S transit thus interrupting mitosis; further failure by cdk2 phosphorylation of E2F would prevent formation of pre-initiation complexes and result in growth arrest (19). However, we note that growth arrest did not occur in vitro and that the in vivo tumors in which the size of the tumor remained the same with GIP treatment might reflect growth arrest. Yet, this growth arrest was released within 7 days. We suggest that irreversible growth arrest did not occur because the GIPs, though capable of pushing transformed cells toward normal homeostasis, are not sufficient to resolve the pathology, which possibly requires cell death. It might be relevant that we observed that transient overexpression of RAR-ß in E:WM-115 melanoma cells resulted in frank apoptosis within 48 h. This was enhanced by the exposure of the transfected melanoma cells to 9 cis-retinoic acid, the RAR–RXR ligand (45) (Figure 10). Further, this observation minimally suggests that the GIPs can identify useful molecular targets, for example RAR-ß (Figures 4A and 7).


Figure 10
View larger version (16K):
[in this window]
[in a new window]
  		Fig. 10. Over-expression of RAR-ß in E:WM-115 cells. Cells were transfected with RAR-ß or empty vector. White, attached live cells; Black, attached rounded cells; Gray, floating dead cells. Vehicle/empty vector, RAR-ß transfected; RAR-ß transfected + RAR-ß ligand[9-cis retinoic acid (9cRA)].

 
Interestingly, our results indicate that while H-TIG-31–164, H-Ha-Rev107-11–162, and H-LRAT1–230 share this 12-amino acid sequence homology and a degree of structural homology within the homologous sequence (Figure 8), biological activity (as cell growth inhibitors) is common only to wild-type peptides from the full-length tumor suppressor proteins and not wild-type LRAT160–171 peptide. However, a single amino acid substitution (Figure 1) obtained the desired result (Figure 2A). Moreover, as discussed in the Molecular modeling Section, while we do not yet fully understand the significance of the leucine residue, we do observe that the unique tri-peptide sequences with growth inhibitory activity have a leucine in Position 3 as does the tri-peptide sequence NYL found in the M4 region of DNA-binding zinc fingers and that the GIP with tri-peptides which have two residues in common with the zinc-finger NYL tri-peptide possess higher growth inhibitory activity. Thus we note that H-TIG3111–123 (NEL) > H-LRAT160–171:C168L (TYL) > H-Rev107111–123 (AQL) > H-LRAT160–171 (TYC) = H-TIG3111–123:L120C (NEC). In terms of single amino acid substitutions and LRAT activity/function, it has been reported that single amino acid substitutions at Position C168 in H-LRAT are to a lesser extent important for LRAT catalysis than C161 mutations and that C168 could be important for structure maintenance or a second transfer reaction not absolutely essential to catalysis (7). Indeed, we have suggested that perhaps mutations at this position have as a consequence additional functions of the LRAT protein, e.g. growth inhibitory (16). Insofar as the significance of the leucine at this position in both the mutated LRAT peptide as well as the wild-type peptides from the full-length tumor suppressor proteins is unknown, we suggest that our work is the first to implicate this leucine residue as important in peptide–DNA binding kinetics and ostensibly the full-length proteins. It is interesting to speculate that AQ, NE, TY in the unique tri-peptides might affect affinity/avidity of binding (Figure 8B). A more complete understanding might be had by the study of binding kinetics of the peptides/full-length proteins with DNA. In light of our observations and recent studies that report another isoform of H-Ha-Rev107 tumor suppressor with acyltransferase properties (8,41), it is interesting to consider that there may exist isoforms of the LRAT enzyme that possess tumor suppressor properties.

In support of our peptide–DNA binding observations (Figure 7), a search in VAST generated several structures that possess partial structural homology to truncated H-LRAT29–200(truncated LRAT). Among these structures is 1BRN, a short segment of DNA complexed with RNase/endonuclease. Indeed, 1BRN barnase interacts with the phosphates of the DNA (47), an interaction we hypothesize occurs through the common tri-peptide RYG in the growth inhibitory peptides. As noted earlier this shared tri-peptide may have relevance in the observed peptide–DNA binding since these residues are also found in RARs and are shown to interact with the phosphate backbone of DNA. We also note that this truncated form of LRAT includes the sequence homologous region and find it interesting that the truncated LRAT is also used to study LRAT enzyme function/activity.

We would like to propose that the sequence homologous peptides mimic behavior of their respective regions within the full-length proteins as suggested by our model (Figure 9). This is a reasonable consideration because our molecular modeling, in which multiple secondary-structure prediction algorithms were applied, suggests an alpha-helix formation of these regions in their respective full-length proteins (data not shown) as well as for the sequence homologous peptides (Figure 8A). Additionally, Duecher et al. (46) demonstrated full-length TIG-31–164 having peri-nuclear localization and suppressing colony formation 50–70%, while our data show that the sequence homologous peptides, e.g. TIG111–123, can localize to the nucleus and inhibit cell proliferation (Figures 1, 3 and 5).

Thus the fact that we find the GIPs localized to the nucleus not only suggests the peptides' behavior can lend insight into the behavior of their full-length proteins, but it also implicates DNA interaction as a functional aspect of these proteins. Of course, these observations and conclusions do not preclude the possibility that these full-length proteins can have additional functions in the cytoplasm and nucleus. However, this is the first study to demonstrate DNA interaction of peptides from these LRAT-like proteins.


   Acknowledgments
 
We thank Susan Garfield and Stephen Wincovitch, Sr of the Confocal Core, Laboratory of Experimental Carcinogenesis, National Cancer Institute, for their technical expertise. ‘Molecular modeling’ was funded by Federal funds from the National Cancer Institute, National Institutes of Health, under Contract No. N01-C0-12400.

Conflict of Interest Statement: None declared.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 DISCUSSION
 References
 

  1. Myers,M.P., Stolavrov,J.P., Eng,C., Li,J., Wang,S.I., Wigler,M.H., Parsons,R. and Tonks,N.K. (1997) P-TEN, the tumor suppressor from human chromosome 10q23, is a dual-specificity phosphatase. Proc. Natl Acad. Sci. USA, 94, 9052–9057.[Abstract/Free Full Text]
  2. Lind,T., Tufaro,F., McCormick,C., Lindahl,U., Lidholt,K. (1998) The Putative Tumor Suppressors EXT1 and EXT2 Are glycosyltransferases required for the biosynthesis of heparan sulfate. J. Biol. Chem., 273, 26265–26268.[Abstract/Free Full Text]
  3. Yokoyama,F., Suzuki,N., Haruki,M., Nishi,N., Oishi,S., Fujii,N., Utani,A., Kleinman,H.K. and Nomizu,M. (2004) Cyclic peptides from the loop region of the laminin alpha 4 chain LG4 module show enhanced biological activity over linear peptides. Biochemistry, 43, 13590–13597.[Medline]
  4. Vidal,M. Liu,W.Q., Gril,B., Assayag,F., Puopon,M.F. and Garbay,C. (2004) Design of new anti-tumor agents interrupting deregulated signaling pathways induced by tyrosine proteins. Inhibition of protein–protein interaction involving Grb2. J. Soc. Biol., 198, 133–137.[Medline]
  5. Shadidi,M. and Sioud,M. (2003) Selective targeting of cancer cells using synthetic peptides. Drug Resist. Updat., 6, 363–371.[CrossRef][Web of Science][Medline]
  6. Haro,I. and Gomara,M.J. (2004) Design of synthetic peptide constructs for the vaccine development against viral infections. Curr. Protein. Pept. Sci., 5, 425–433.[CrossRef][Web of Science][Medline]
  7. Mondal,M.S., Ruiz,A., Bok,D. and Rando,R.R. (2000) Lecithin retinol acyltransferase contains cysteine residues essential for catalysis. Biochemistry, 39, 5215–5220.[CrossRef][Medline]
  8. Anantharaman,V. and Aravind,L. (2003) Evolutionary History Structural Features and Biochemical Diversity of the NipC/P60 Superfamily of Enzymes. Genome Biol., 4, R 11–R 12.[CrossRef][Medline]
  9. Ruiz,A., Kuehn,M.H., Andorf,J.L., Stone,E., Hageman,G.S. and Bok,D. (2001) Genomic organization and mutation analysis of the gene encoding lecithin retinol acyltransferase in human retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci., 42, 31–37.[Abstract/Free Full Text]
  10. Hajnal,A., Klemenz,R. and Schafer,R. (1994) Subtraction cloning of H-rev107, a gene specifically expressed in H-ras resistant fibroblasts. Oncogene, 9, 479–490.[Web of Science][Medline]
  11. Bastien,J. and Rochette-Egly,C. Nuclear retinoid receptors and the transcription of retinoid-target genes. Gene, 2004, 328, 1–16.[CrossRef][Web of Science][Medline]
  12. Guo,X., Knudsen,B.S., Peehl,D.M., Ruiz,A., Bok,D., Rando,R.R., Rhim,J.S., Nanus,D.D. and Gudas,L.J. (2002) Retinol metabolism and lecithin:retinol acyltransferase are reduced in cultured human prostate cancer cells and tissue specimens. Cancer Res., 62, 1654–1661.[Abstract/Free Full Text]
  13. Guo,X., Ruiz,A., Rando,R., Bok,D. and Gudas,L. (2000) Esterification of all-trans-retinol in normal human epithelial cell strains and carcinoma lines from oral cavity, skin and breast: reduced expression of lecithin:retinol acyltransferase in carcinoma lines. Carcinogenesis, 21, 1925–1933.[Abstract/Free Full Text]
  14. Andreola,F., Giandomenico,V., Spero,R. and De Luca,L.M. (2000) Expression of a smaller lecithin:retinol acyl transferase transcript and reduced retinol esterification in MCF-7 cells. Biochem. Biophys. Res. Commun., 279, 920–924.[CrossRef][Web of Science][Medline]
  15. Kurlandsky,S., Duell,E., Kang,S., Voorhees,J. and Fisher,G. (1996) Auto-regulation of retinoic acid biosynthesis through regulation of retinol esterification in human keratinocytes. J. Biol. Chem., 271, 15346–15352.[Abstract/Free Full Text]
  16. Simmons,D.P., Andreola,F. and De Luca,L.M. (2002) Human melanomas of fibroblast and epithelial morphology differ widely in their ability to synthesize retinyl esters. Carcinogenesis, 23, 1821–1830.[Abstract/Free Full Text]
  17. Sers,C., Emmenegger,U., Hussman,K., Bucher,K., Andres,A.C. and Schäfer,R. (1997) Growth-inhibitory activity and downregulation of the class ii tumor- suppressor gene H-rev107 in tumor cell lines and experimental tumors. J. Cell Biol., 136, 935–944.[Abstract/Free Full Text]
  18. Sun,S., Wan,H., Yue,P., Hong,W. and Lotan,R. (2000) Evidence that retinoic acid receptor beta induction by retinoids is important for tumor cell growth inhibition. J. Biol. Chem., 275, 17149–17153.[Abstract/Free Full Text]
  19. Aguda,B.D. and Tang,Y. (1999) The kinetic origins of the restriction point in the mammalian cell cycle. Cell Prolif., 32, 321–335.[CrossRef][Web of Science][Medline]
  20. Santoni-Rugiu,E., Falck,J., Mailand,N., Bartek,J. and Lucas,J. (2000) Involvement of Myc activity in a G1/S-promoting mechanism parallel to the pRb/E2F pathway. Mol. Cell. Biol., 20, 3497–3509[Abstract/Free Full Text]
  21. Clark,W.H. (1991) Tumor progression and the nature of cancer. Brit. J. Cancer., 64, 631–644.[Medline]
  22. Ahmedin,J., Tiwari,R.C., Murray,T., Ghafoor,A., Samuels,A., Ward,E., Feuer,E.J. and Thun,M.J. (2004) Cancer statistics, CA Cancer J. Clin., 54, 8–29.[Abstract/Free Full Text]
  23. Rigel,D., Friedman,R. and Kopf,A. (1996) Lifetime risk for development of skin cancer in the U.S. population: Current estimate is now 1 in 5. J. Am. Acad. Dermatol., 35, 1012–1013.[CrossRef][Web of Science][Medline]
  24. Weber,P.J., Bader,J.E., Folkers,G. and Beck-Sickinger,A.G. (1998) A fast and inexpensive method for N-terminal fluorescein labeling of peptides. Bioorg. Med. Chem. Lett., 8, 597–600.[CrossRef][Medline]
  25. Hu,Z., Sun,Y. and Garen,A. (1999) Targeting tumor vasculature endothelial cells and tumor cells for immunotherapy of human melanoma in a mouse xenograft model. Proc. Natl Acad. Sci. USA, 96, 8161–8166.[Abstract/Free Full Text]
  26. Albrecht,M., Tosatto,S.C.E., Lengauer,T. and Valle,G. (2003) Simple consensus procedures are effective and sufficient in secondary structure prediction. Protein Engg, 16, 459–462.
  27. Jorgensen,W.L., Maxwell,D.S. and Tirado-Rives,J. (1996) Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc., 118, 11225–11235.[CrossRef][Web of Science]
  28. Still,W.C., Tempczyk,A., Hawley,R.C. and Hendrickson,T. (1990) Semianalytical treatment of solvation for molecular mechanics and dynamics. J. Am. Chem. Soc., 112, 6127–6129.[CrossRef][Web of Science]
  29. Wimmel A, Wiedenmann,B. and Rosewicz,S. (2003) Autocrine growth inhbition by transforming growth factor beta-1(TGFbeta-1) in human neuroendocrine tumor cells. Gut, 52, 1308–1316.[Abstract/Free Full Text]
  30. Reeves,R., Edberg,D. and Li,Y. (2001) Architectural transcription factor HMGI (Y) promotes tumor progression and mesenchymal transition of human epithelial cells. Mol. Cell Biol., 21, 575–594.[Abstract/Free Full Text]
  31. Balmer,J. and Blomhoff,R. (2002) Gene expression regulation by retinoic acid. J. Lipid Res., 43, 1773–1808.[Abstract/Free Full Text]
  32. Toulouse,A. (2000) RARß involvement in enhancement of lung tumor cell immunogenecity revealed by array analysis. FASEB J., 14, 1224–1232.[Abstract/Free Full Text]
  33. Harwell,R.M., Porter,D.C., Danes,C. and Keyomarsi,K. (2000) Processing of cyclin E differs between normal and tumor breast cells. Cancer Res., 60, 481–489.[Abstract/Free Full Text]
  34. Bannister,A. and Kouzarides,T. (1992) Basic peptides enhance protein/DNA interaction in vitro. Nucleic Acids Res., 20, 3523.[Free Full Text]
  35. Spanjaard,R.A., Sugawara,A., Ikeda,M. and Chin,W.W. (1995) Evidence that retinoid receptors mediate retinoid-dependent transcriptional activation of the retinoic acid receptor beta gene in S91 melanoma cells. J. Biol. Chem., 270, 17429–17436.[Abstract/Free Full Text]
  36. McKenna,N. and O'Malley,B. (2001) Nuclear receptors, coregulators, ligands, and selective receptor modulators: making sense of the patchwork quilt. Ann. NY Acad. Sci., 949, 3–5.[CrossRef][Web of Science][Medline]
  37. McKenna,N. and O'Malley,B. (2002) Minireview: nuclear receptor coactivators—an update. Endocrinol., 143, 2461–2465.[Abstract/Free Full Text]
  38. Ellenberger,T.E., Brandl,C.J., Struhl,K. and Harrison,S.C. (1992) The GCN4 basic leucine zipper DNA as a dimmer of uninterrupted alpha helices: crystal structure of the protein–DNA complex. Cell, 71, 1223–1237.[CrossRef][Web of Science][Medline]
  39. Hsieh,J.-C, Whitfield,G.K., Jurutka,P.W., Haussler,C.A., Thatcher,M.L., Thompson,P.D., Dang,H.T.L., Galligan,M.A., Oza,A.K. and Haussler,M.R. (2003) Two basic amino acids C-terminal of the proximal box specify functional binding of the vitamin D receptor to its rat osteocalcin deoxyribonucleic acid responsive elelment. Endocrinol., 144, 5065–5080.[Abstract/Free Full Text]
  40. Rastinejad,F., Wagner,T., Zhao,Q. and Khorasanizadeh,S. (2000) Structure of the RXR–RAR DNA-binding complex on the retinoic acid response element DR1. EMBO J., 19, 1045–1054.[CrossRef][Web of Science][Medline]
  41. Jahng,W.J., Xue,L. and Rando,R.R. (2003) Lecithin retinol acyltransferase is a founder member of a novel family of enzymes. Biochemistry, 42, 12805–12812.[CrossRef][Medline]
  42. Sherr,C.J. and Roberts,J.M. (2004) Living with or without cyclins and cyclin-dependent kinases. Genes and Dev., 18, 2699–2711.[Abstract/Free Full Text]
  43. Matsumoto,Y. and Maller,J.L. (2004) A centrosomal localization signal in cyclin E required for cdk2-independent S phase entry. Science, 5697, 885–888.
  44. Laing,J.H.E., Wilson,G.D. and Martindale,C.A. (2003) Proliferation rates in human malignant melanoma:relationship to clinicopathological features and outcome. Melanoma Res., 13, 271–277.[CrossRef][Web of Science][Medline]
  45. Heyman,R.A., Mangelsdorf,D.J., Dyck,J.A., Stein,R.B., Eichele,G.l; Evans,R.M. and Thaller,C. (1992) 9-cis Retinoic acid is a high affinity ligand for the retinoid X receptor. Cell, 68, 397–406.[CrossRef][Web of Science][Medline]
  46. Deucher,A., Nagpal,S., Chandraratna,R.A., Sepio,D., Robinson,N.A., Dashti,S.R. and Eckert,R.L. (2000) The carboxy-terminal hydrophobic domain of TIG3, a class II tumor suppressor protein, is required for appropriate cellular localization and optimal biological activity. Int. J. Oncol., 17, 1195–1203.[Web of Science][Medline]
  47. Buckle,A.M. and Fersht,A.R. (1994) Subsite binding in an Rnase: structure of a barnase-tetranucleotide complex at 1.76-Å resolution. Biochem., 33, 1644–1653.[CrossRef][Medline]
Received March 1, 2005; revised May 9, 2005; accepted September 28, 2005.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
27/4/693    most recent
bgi235v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (3)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Simmons, D. P.
Right arrow Articles by De Luca, L. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Simmons, D. P.
Right arrow Articles by De Luca, L. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?