Phospholipid- and GTP-dependent activation of cholera toxin and phospholipase D by human ADP-ribosylation factor-like protein 1 (HARL1).

ADP-ribosylation factors (ARFs), 20-kDa guanine nucleotide-binding proteins named for their ability to activate cholera toxin (CT) ADP-ribosyltransferase activity, have a critical role in vesicular transport and activate a phospholipase D (PLD) isoform. Although ARF-like (ARL) proteins are very similar in sequence to ARFs, they were initially believed not to activate CT or PLD. mRNA for human ARL1 (hARL1), which is 57% identical in amino acid sequence to hARF1, is present in all tissues, with the highest amounts in kidney and pancreas and barely detectable amounts in brain. Relative amounts of hARL1 protein were similar to mRNA levels. Purified hARL1 (rARL1) synthesized in Escherichia coli had less activity toward PLD than did rARF1, although PLD activation by both proteins was guanosine guanosine 5'-(gamma-thio)triphosphate (GTPgammaS)-dependent. ARL1 stimulation of CT-catalyzed ADP-ribosylation was considerably less than that by rARF1 and was phospholipid dependent. GTPgammaS-binding by rARL1 was also phospholipid- and detergent-dependent, and in assays containing phosphatidylserine, was greater than that by rARF1. In vitro, the activities of rARL1 and rARF1 are similar. Rather than being a member of a separate subfamily, hARL1, which activates PLD and CT in a phospholipiddependent manner, appears to be part of a continuum of ARF family proteins.

ADP-ribosylation factors (ARFs), 1 a family of 20-kDa GTPbinding proteins, were discovered as activators of cholera toxin (CTA)-catalyzed ADP-ribosylation of G s ␣ (1) and subsequently shown to serve as allosteric activators of the toxin catalytic unit (2). ARFs are known to function in vesicular transport and membrane trafficking (3) and, more recently, were found to activate phospholipase D (PLD), an enzyme that cleaves phosphatidylcholine (PC) to produce phosphatidic acid and choline (4,5). ARFs require phospholipid and/or detergent for high affinity GTP binding and optimal enhancement of CTA activity (6,7). There are six known mammalian ARF proteins, which can be divided into three classes based on deduced amino acid sequence, protein size, and gene structure (8,9). Members of all three classes can associate with specific intracellular membranes, consistent with roles in vesicular transport, and can also activate PLD (10), processes that could be intimately related.
Although the ARF-like proteins (ARLs) have deduced amino acid sequences very similar to ARFs (11)(12)(13)(14)(15), they appeared initially not to activate cholera toxin and failed to rescue the lethal Saccharomyces cerevisiae mutant (11) in which the arf1 and arf2 genes are no longer functional. Like ARFs, ARLs bind GTP; ARLs, however, hydrolyze bound GTP, whereas ARFs do not (16). Some ARLs exhibit tissue-, cell-and/or differentiation stage-specific expression (12)(13)(14)(15). ARL1 reportedly localized to the Golgi in normal rat kidney cells (17) and in S. cerevisiae (18), suggesting a role in vesicular trafficking. Unlike the combination of ARFs 1 and 2, the yARL1 gene knock out construct was not lethal. Moreover, yARL1 did not rescue the arf1 Ϫ arf2 Ϫ construct and thus may not have a role in constitutive secretion, but rather in regulated secretion.
We report here that the hARL1 protein is abundant in most human tissues, except brain (adult and fetal). Human ARL1, which is more similar in amino acid sequence to hARF1 than to other human ARLs, can, under certain conditions, activate PLD and enhance CTA-catalyzed ADP-ribosyltransferase activity.

Expression and Purification of Recombinant Proteins-
The open reading frame of human ARL1 was obtained by polymerase chain reaction using primers that incorporated NdeI and BamHI sites at the initiating methionine and 6 base pairs downstream of the stop codon, respectively. For the His-tagged hARL1 fusion protein, the polymerase chain reaction product was inserted into vector pET15b (Novagen), yielding pET15bhARL1. For the nonfusion protein, the polymerase chain reaction product was inserted into vector pT7 (20), yielding pT7hARL1. BL21(DE3) cells transformed with the above constructs were grown to a density of A 600 ϳ 1.0 in LB broth containing ampicillin (100 g/ml), at which time protein synthesis was induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside. After 3 h, cells were harvested by centrifugation, washed once in 20 mM Tris, pH 7.4, 1 mM EDTA, and stored at Ϫ80°C. For large scale protein production, one liter of LB broth containing ampicillin was inoculated with 5 ml of an overnight culture, followed by shaking at 37°C. When the culture reached A 600 of ϳ 0.6 -0.8, protein production was induced with 0.5 mM isopropyl-1thio-␤-D-galactopyranoside. After 3 h, bacteria were collected by centrifugation and stored at Ϫ20°C. Cell pellets were suspended in 10 ml of phosphate-buffered saline, pH 7.4, containing lysozyme (0.5 mg/ml) and disrupted by sonification. The lysate was centrifuged for 30 min at 100,000 ϫ g and the supernatant recovered. His-tagged protein was isolated on Ni 2ϩ -nitrilotriacetic acid resin (Qiagen, Chatsworth, CA) by standard methods; nonfusion protein was isolated by gel permeation chromatography (20). Purity was assessed by Coomassie Blue staining of SDS-polyacrylamide gels, and protein was quantified by Coomassie Blue dye-binding assays (Bio-Rad).
PLD Assays-DPPC hydrolysis was assayed as described previously

FIG. 2. Reactivity of anti-hARL1 polyclonal antibodies with
ARFs and ARLs detection of hARL1 protein in human tissues. A, recombinant nonfusion ARFs as well as nonfusion and 6-His ARLs (2 g each) (B) were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes, reacted with anti-hARL1 polyclonal antibodies, developed, and exposed to Hyper-film-MP. Positions of protein standards (kDa) are noted in A. C, a multiple human tissue blot (MTN blot, 6.5-100-kDa proteins, purchased from CLONTECH) was incubated with a rabbit polyclonal antibody against hARL1. The chemiluminescent signal was imaged on Hyper-film-MP with a 5-min exposure. hARL1 was detected, as indicated by the arrows. Primary and secondary antibodies and luminol substrate were removed from the blot using the multiple tissue Western blot-stripping protocol, before it was reacted with a rabbit anti-␣-tubulin antiserum. Positions of protein standards (kDa) are to the left. nitrocellulose filters and washed (five times) with 2 ml of buffer; radioactivity was quantified by liquid scintillation counting of dry filters.
Northern Blotting-Filters with mRNA from human tissues (CLON-TECH) were hybridized as described (23) with a 3Ј-end-32 P-labeled, 45-base oligonucleotide probe (5Ј-CCG AGT TCC AAA CAG ACT GGA AAA TAT ACT TGA GAA AAA GCC ACC-3Ј) complementary to the 5Ј-end of hARL1. The probe (specific activity: ϳ6 ϫ 10 8 cpm/mg) was prepared by incubation with [␣-32 P]dATP and terminal deoxynucleotidyltransferase, followed by purification on a NAP-5 column (Amersham Pharmacia Biotech, Uppsala, Sweden). Blots were also hybridized with a random-primed, 32 P-radiolabeled human actin probe as an internal control. Bands were quantified on a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) for calculation of ARL/actin ratios.
Polyclonal Antibody Production-Purified recombinant His-tagged hARL1, prepared as described above, was further purified by SDS-PAGE. A band corresponding to the pure protein was excised from the gel for immunization of rabbits as described (24). The preparation of antibodies against hARL5 has been described (18).

TABLE II Effect of nucleotides on rARL1 stimulation of CTA activity
Assays contained 6 g of the indicated protein (with or without 200 M GTP␥S or GDP) and were carried out as described under "Experimental Procedures." rARL1-D1 and rARL1-D2 were two separate preparations of rARL1 dialyzed against 7 M urea in KEND buffer (10 mM potassium phosphate, pH 7.4, 1 mM EDTA, 100 mM NaCl, 1 mM dithiothreitol) (to remove bound nucleotide) for 24 and 48 h, respectively followed by several changes of buffer without urea (20   antibodies against rARL1; reactivity was detected with ECL. As a control, primary and secondary antibodies were removed from the blot using the multiple tissue Western blot-stripping protocol and the blot reacted with rabbit anti-␣-tubulin antiserum.

RESULTS AND DISCUSSION
Using a hARL1-specific oligonucleotide probe for Northern blotting, hARL1 mRNA was detected in most human tissues (lung, heart, liver, placenta, kidney, pancreas, and skeletal muscle) with the largest amount in pancreas and kidney; less ARL1 mRNA was present in adult or fetal brain (Fig. 1). The hARL1-specific oligonucleotide probe hybridized with three distinct bands of 3.2, 1.7, and 1.2 kb in RNA from all tissues except brain, where only the 3.2-kb species was found. In additional blotting experiments, a different sequence-specific probe yielded identical results for hARL1 (data not shown). The level of rat ARL1 mRNA was reported to be reduced after differentiation of 3T3-L1 cells and only a single 2.2-kb mRNA was detected (14). In contrast, Lowe et al. (17) reported that the major rat ARL1 mRNA had a size of 1.8 kb and was present in all tissues tested, albeit at relatively low levels in lung and spleen. The discrepancy in size between human and rat message is probably due to species differences. In addition to those shown for rat ARL1 (17), multiple mRNA species were also noted for human ARL3 (13) and rat ARL5 (15).
Initially, hARL1 was thought not to activate PLD significantly (21,26). Activation of a highly purified PLD by hARL1 was less than that by hARF1 (Fig. 4), but was GTP␥S-dependent (Fig. 5). Activation of PLD by hARL, along with the prior observation of rat ARL1 localization to the Golgi, could be consistent with a role in vesicular trafficking similar to that of ARFs.
Based on protein size, deduced amino acid sequence, phylogenetic analysis, and gene structure, the six mammalian ARFs can be grouped into three classes (8). ARL proteins are structurally related to ARFs, but due to their relative ineffectiveness in the activation of cholera toxin and inability to rescue the lethal arf1Ϫ arf2Ϫ double deletion mutant in yeast (8), they were not classified as ARFs. Alignment of the amino acid sequence of hARL1 with those of hARFs and other hARLs reveals the greatest degree of identity exists between hARL1 and class II ARFs (Table IV). Sequences of hARF4 and hARF5 (180 amino acids) are 58% identical to hARL1. Deduced amino acid sequences of hARL1 and hARF1 are 57% identical, and the functional consequences of some of these differences have been demonstrated (21). hARL1 is only 35-43% identical to the other four known ARLs. Based on this alignment, it appears that human ARLs may fall into classes in a manner analogous to human ARFs (Fig. 6). ARL1 differs on one hand from ARLs 2 and 3, which are 53% identical and on the other from ARLs 4 and 5, which are 60% identical. ARL2 is, however, only 27 and 29% identical to ARLs 4 and ARL5, respectively, and ARL3 is 37 and 35% identical to ARL4 and ARL5, respectively. We, therefore, propose that there are at least three distinct classes of ARL (I, II, and III); human ARL1 (Class I) appears to be different, from ARLs 2 and 3 (Class II) and ARLs 4 and 5 (Class III). Human ARLs 2 and 3 did not activate PLD, although, under the same assay conditions, they bound significantly more GTP␥S than did hARL1. 2 Because its sequence is more similar to those of hARFs than to other hARLs, and in light of its activity toward PLD and CTA in the presence of appropriate lipids, hARL1 may be part of an ARF continuum rather than a member of a separate subfamily and may have physiological functions similar to those of ARFs.