PTEN Redundancy: Overexpressing lpten, a Homolog of Dictyostelium discoideum ptenA, the Ortholog of Human PTEN, Rescues All Behavioral Defects of the Mutant ptenA−

Mutations in the tumor suppressor gene PTEN are associated with a significant proportion of human cancers. Because the human genome also contains several homologs of PTEN, we considered the hypothesis that if a homolog, functionally redundant with PTEN, can be overexpressed, it may rescue the defects of a PTEN mutant. We have performed an initial test of this hypothesis in the model system Dictyostelium discoideum, which contains an ortholog of human PTEN, ptenA. Deletion of ptenA results in defects in motility, chemotaxis, aggregation and multicellular morphogenesis. D. discoideum also contains lpten, a newly discovered homolog of ptenA. Overexpressing lpten completely rescues all developmental and behavioral defects of the D. discoideum mutant ptenA−. This hypothesis must now be tested in human cells.


Introduction
PTEN is one of the most commonly mutated tumor suppressor genes in the progression of human cancers [1][2][3]. It has been demonstrated to play a prominent role in several cellular behaviors, including basic cell motility, chemotaxis and invasion [4][5][6][7][8][9][10]. PTEN functions as a phosphatase that regulates the signal transduction molecule phosphatidylinositol-3, 4, 5-triphosphate (PIP 3 ) [11]. There are three homologs of the PTEN gene in the human genome [12][13][14][15][16]. In addition, Poliseno et al. (2010) [17,18] found that a PTEN pseudogene, PTENP1, regulates the level of PTEN protein and acts as a growth suppressor. The presence of PTEN homologs in the human genome, therefore, raises the possibility that one of them may be able to substitute functionally for a mutated PTEN under inducing conditions, thus suppressing tumorigenesis, a possibility heretofore not tested.
The amoeba Dictyostelium discoideum, an exceptional model for studying the regulation of human cell motility and chemotaxis [19][20][21][22][23][24][25][26][27], contains the gene ptenA, an ortholog of the human PTEN gene. Deletion of ptenA in D.discoideum causes major defects in lateral pseudopod suppression, motility, chemotaxis and natural aggregation [28][29][30][31][32][33][34]. As is the case for human PTEN, PtenA in D.discoideum dephosphorylates phospahtidylinositol (3,4,5)-trisphosphate (PIP 3 ) to form phophatidylinositol (4,5)-bisphosphate (PIP 2 ) [35,36] and mediates PIP 3 oscillations [37][38][39][40][41], which correlate with actin polymerization and pseudopod extension [30,[39][40][41][42][43]. PtenA was originally thought to be the sole phosphatase for the dephosphorylation of PIP 3 to PIP 2 in D. discoideum. However, after global stimulation of ptenA 2 cells with the chemoattractant cAMP, the concentration of PIP 3 increases, but then declines [36], indicating that PIP 3 is degraded to PIP 2 in the absence of PtenA, presumably by another phosphatase. Moreover, Hoeller and Kay [32] demonstrated that when suspensions of ptenA 2 cells were pulsed with cAMP to induce chemotactic responsiveness, they were able to undergo efficient chemotaxis. However, unlike earlier studies in which the concentration of the cAMP gradient, generated in vitro was in the range of that estimated for the gradient in the front of a natural cAMP wave [44], Hoeller and Kay [32] employed a cAMP gradient generated in a concentration range 10 times higher than that employed in the prior studies of ptenA 2 chemotaxis [29,30] and, therefore, 10 times higher than that estimated for the natural cAMP wave that induces chemotaxis in natural populations [44]. The studies of PIP 3 degradation in ptenA 2 cells after global cAMP stimulation [36] and chemotaxis of ptenA 2 cells in high cAMP   RT-PCR revealed that lpten is up-regulated during aggregation. lpten expression during development was assessed by RT-PCR and quantified by densitometry. lpten expression was up-regulated more than 10 fold. P1and P2 (see Table S1) demark the positions of the primers used to amplify the 300bp lpten cDNA fragment (F). RT-PCR of the large subunit ribosomal RNA, rnlA, was assessed for comparability of gel loading. C. Scheme for lpten disruption. The positions of primers P3 to P7 (see Table S1) are demarcated for amplification of lpten in control Ax2 and lpten 2 cells, to generate the undisrupted lpten genomic fragment F1, the disrupted lpten genomic fragment of lpten 2 , F2 genomic fragment F2, and a partial lpten 2 genomic fragment with a partial bsr cassette, F3. D. Verification of lpten 2 disruption by PCR. See panel C for the positions of the primers to generate fragments F1, F2 and F3. E. Verification that the lpten transcript is missing in the lpten 2 mutant using RT-PCR with the primers LptenFW and PtencDNArv for ptenA, and PtenAcDNAFW and ptencDNArv to demonstrate the presence of ptenA in lpten 2 . See Table S1 for description of primers. F. The completion of multicellular morphogenesis by the formation of fruiting bodies in control Ax2 cells. G. The absence of morphogenesis by ptenA 2 cells. H. The completion of multicellular morphogenesis by lpten 2 cells. doi:10.1371/journal.pone.0108495.g001 concentration gradients [32], suggested to us that there might be an alternative PIP 3 phosphatase that could substitute for ptenA.
We therefore searched the D.discoideum database (http:// dictybase.org/) and found a second ortholog of human PTEN and homolog of ptenA [28,29], which we named lpten because it contained unique LIM domains. Here we show that cells of the lpten deletion mutant, lpten 2 , exhibit defects in behavior similar to those in ptenA 2 cells, but the defects are far weaker. To test for redundant function, we overexpressed lpten in a ptenA 2 background. Overexpression resulted in the complete normalization of the defective behaviors of ptenA 2 cells. The ptenA 2 defects that were normalized included the following: abnormal aggregation, the absence of multicellular morphogenesis, the loss of lateral pseudopod suppression, increased turning, decreased cellular velocity, aberrant chemotaxis in a cAMP gradient generated in the standard concentration range and aberrant natural aggregation. We further show that pulsing ptenA 2 cells with cAMP, which induces chemotactic competency in a high cAMP concentration gradient [32], is accompanied by up-regulation of lpten expression. We therefore conclude that lpten plays a similar, but less prominent in pseudopod suppression, motility and chemotaxis role than its homolog ptenA, but when overexpressed in the ptenA 2 mutant, rescues all of the ptenA 2 defects. This raises the question of whether any of the homologs of human PTEN might also be induced to function redundantly in cancer cells carrying mutations in PTEN.
DNA, RNA purification, cloning and sequencing Isolation, purification, amplification and sequencing of all the genomic DNA, RNA and cDNA fragments from D. discoideum Ax2, mutant strains and plasmids was done as previously described [48]. Plasmids and competent cells were obtained from Life Technologies, (Carslbad CA, USA) [48]. For RNA, recombinant RNasin Ribonuclease (Promega, Madison, WI, USA) was added to inhibit RNA degradation. RNA was additionally purified from residual genomic DNA by using RNAeasyPlus (Qiagen, Ventura, CA, USA) according to the manufacturer's instructions. The primers used in this study are listed in a Table S1. Transformants were generated as described [48] and selected for using either 10 mg/ml Blasticidin S (Enzo Life Science, Farmingdale, NY, USA) and/or 20-70 mg/ml G418 (Sigma-Aldrich, St.Louis, USA). For clonal growth selection of the ptenA 2 /lpten oe and lpten 2 / lpten oe strains, cells were sorted by FACS as described [48].

Generation of a Dictyostelium discoideum knock out strain
A plasmid was generated that contained a bsr resistance cassette containing the gene coding for blasticidin deaminase flanked by lpten genomic fragments, as described by Torija et al [52] and diagrammed in Figure 1. In brief, a genomic fragment (F1) ( Figure 1C) containing the lpten open reading frame and upstream and downstream regions was cloned and incubated in the presence of a PvuII-digested EZTN-plasmid (Epicenter, Madison, WI, USA), which contained a transposon bearing the Blasticidin S resistance marker [53]. Transposase was used to insert the transposon carrying the bsr gene into the lpten-containing plasmid (Epicentre). Transformed bacteria were selected for by tetracyclin (15 mg/ml) and kanamycin (50 mg/ml). Bacterial colonies bearing the plasmid that had a transposon within the genomic fragment were identified using primer M13, T7 and EZTN-R (Table S1). For D. discoideum Ax2 [32,45] transformation, a fragment of the plasmid carrying the insert containing the bsr-resistance cassette close to the 59 end of the lpten coding region was amplified as described by Torija et al. [52], except for the use of Expand Long Template PCR Polymerase (Roche, Indianapolis, IN, USA) [48]. Selection was done with increasing Blasticidin S concentrations. Surviving cells were clonally plated on nutrient plates in the presence of 30 mg/ml Blastidicin S and Klebsiella aerogenes. Colonies were harvested using MasterAmp Buccal Swab DNA extraction solution (Epicentre) and subjected to D. discoideum colony PCR [48,52].

Generation of lpten overexpression constructs
To obtain an RFP-Lpten fusion, we amplified and cloned the coding region of lpten into the extrachromosomal plasmid pDM354 [54]. Cloning and recombination of the lpten cDNA were performed as described [48]. Transformed bacterial clones were identified by colony PCR.

RNA expression analyses
To study expression levels, RT-PCR was performed using the LongRange 2Step RT-PCR Kit (Qiagen) [48]. 2 mg of total RNA, pretreated at 65uC for 5 min, underwent the reverse transcription reaction in a total volume of 20 ml using OligodT primer supplied by the manufacturer. The resulting cDNA was amplified using the Long Range Expand Polymerase Kit (Roche, Indianapolis, IN) and primer P1 and P2 (Table S1, Figure 1). RNA expression levels were quantified under subsaturation conditions as described [55], using the densitometry function of the 2D-DIAS software program [56].

Analyses of basic cell behavior and chemotaxis
For analyses of basic motility in the absence of chemoattractant and during chemotaxis, cells were harvested from developmental filters at the onset of aggregation, when velocity and chemotactic responsiveness were maximal [57]. For basic behavior in buffer, cells were analyzed on the glass wall of a Sykes-Moore chamber perfused with the buffer BSS according to methods previously described [20,[58][59][60]. The methods for measuring chemotaxis in a Zigmond chamber have also been described in detail elsewhere [60][61][62][63]. For a ''low cAMP concentration'' gradient, the source well contained 1 mM cAMP. For a ''high cAMP concentration'' gradient, the source well contained 10 mM cAMP.

2D-and 3D-DIAS analyses of cell behavior
Cell images were digitally acquired using iStopMotion software (Boinx Software, www.boinx.com) and converted to QuickTime format for edge detection, perimeter reconstruction and motion analysis of cell behavior with 2D-DIAS software [64], as previously described [58,65]. Descriptions of motility and chemotaxis parameters are presented in table S2 [58,64,65]. For 3D reconstruction, cells were optically sectioned and analyzed as previously described [20,58,59,[66][67][68], except that the resulting QuickTime movies of optical sections were exported into jpeg files and converted into JDIAS movies using an up-graded version of 3D-DIAS ( [20,58,69], JDIAS 4.1 (Soll et al. 2014 in prep.). The in-focus perimeters were automatically outlined in each optical section using a pixel complexity algorithm [64,70]. Pseudopods were manually traced.

Analyses of cell behavior in a natural wave
Natural waves were relayed in populations of cells aggregating on a plastic surface in submerged cultures in 35 mm Petri dishes (Fisherbrand, Pittsburg, PA). Cell behavior was recorded and motion analyzed as previously described [30,66,71,72]. Cell behavior in this case was analyzed using JDIAS 4.1.

cAMP-pulsing of cells in suspension
Cells were pulsed as described by Hoeller and Kay [32]. In brief, 2610 7 cells from a growth culture were washed free of nutrients, shaken in a suspension culture for 1h in buffered salts solution, and subsequently pulsed with 80 nM cAMP at 6 min intervals for 6 h, until aggregate formation was visually observed.

Results
Lpten is a homolog of ptenA PtenA of D. discoideum [28,29] contains two functionally important and conserved domains, which are present in human PTEN, a CDC14-dual specificity phosphatase (protein cluster COG2453) [73], found in members of the protein tyrosine phosphatase superfamily, and the lipid-C 2 -binding domain, Pten-C2 (PFAM10409) [74] ( Figure 1A). A database search of the genome sequence of D. discoideum revealed a second ortholog of the human PTEN gene, lpten, (accession number KF430369), which encodes a protein that also contains the conserved CDC14 dual specificity phosphatase domain and the lipid-C2 binding domain ( Figure 1A). In addition, this homolog contains five LIM domains that together contain a total of 38 putative zinc-binding sites ( Figure 1A). Because of the LIM domains, we have named G. Frequency of lateral pseudopod formation in buffer.

Expression of lpten
The ptenA expression was previously shown to be low during D. discoideum growth and to increase during the preaggregative period of development [75,76]. A reverse transcriptase-polymerase chain reaction (RT-PCR) was employed to assess lpten expression in parental Ax2 cells during growth and at the end of the preaggregative period preceding chemotaxis, using a 300 bp probe (F) that spanned exons 3 and 4, as diagrammed in Figure 1B. Lpten was expressed in log phase cells (0 hours) at a very low level and was dramatically up-regulated, approximately ten fold or more in developing cultures at the onset of aggregation (8 hours) ( Figure 1B).

Disruption of lpten and mutant rescue
To disrupt lpten, Ax2 cells were transformed with an integrative construct, as diagrammed in Figure 1C. Integration was confirmed by PCR ( Figure 1D), using primers P5 and P6 to generate fragment F2, as diagrammed in Figure 1C. To further confirm integration, a portion of the disrupted lpten gene was amplified with primers P7 and P6, to generate fragment F3 ( Figure 1C and D). Sequencing of the product F3 confirmed integration. In contrast to ptenA 2 cells, which exhibited a major increase in generation time from 9 to 14 hours [30,31], lpten 2 cells exhibited a generation time of approximately 8 hours, similar to that of the parental strain Ax2 cells. And in contrast to ptenA 2 cells, which do not complete aggregation ( Figure 1G), lpten 2 cells underwent aggregation and multicellular morphogenesis, forming fruiting bodies ( Figure 1H). The lpten 2 mutant was rescued by transformation with a plasmid containing lpten fused to rfp under the regulation of the actin 15 promoter. The complemented mutant strain, lpten 2 /lpten oe , grew with the same generation time as parental Ax2 cells, aggregated and formed fruiting bodies (data not shown). Therefore, deleting lpten resulted in no measurable growth or obvious developmental defect. lpten 2 cells exhibit a minor defect in basic cell motility Using computer-assisted 2D and 3D reconstruction and motion analysis systems, we previously demonstrated that aggregation-competent ptenA 2 cells perfused with a K + -based buffer [30,77] lacking chemoattractant exhibited a 50% decrease in velocity, a major increase in turning and a fourfold increase in lateral pseudopod formation [30]. Using the same computer-assisted methods, we found that lpten 2 cells exhibited basic behavioral defects similar to those of ptenA 2 [30] cells, in instantaneous velocity, percent cells with velocities $9 mm per minute and turning ( Figure 2A). However, the defects although significant (p value ,0.05), were less pronounced. The defects were evident in comparisons of computerreconstructed cell perimeter tracks ( Figure 2C), when compared to those of parental Ax2 cells ( Figure 2B) or complemented lpten 2 /lpten oe cells ( Figure 2D). 3D reconstructions performed with 3D-DIAS software [58,64,70] revealed that lpten 2 cells, like ptenA 2 cells [30], formed lateral pseudopods, which initiate turns, at frequencies higher than Ax2 and lpten 2 /lpten oe ( Figure 2E, G and F, respectively). 2D measurements of the frequency of lateral pseudopods formed by reconstructed control (Ax2) cells, lpten 2 cells and lpten 2 /lpten oe cells, supported this conclusion ( Figure 2G). The frequency of the mutant was over twice that of Ax2 and lpten 2 /lpten oe cells ( Figure 2G).  We previously demonstrated that ptenA 2 cells undergoing positive chemotaxis in a spatial gradient of cAMP generated in the concentration range estimated for the front of a natural wave [44] (i.e., low cAMP concentration gradient), exhibited approximately a 50% decrease in velocity and a similar decrease in directional persistence [30]. Moreover, the efficiency of chemotaxis, measured by the chemotactic index (CI), was reduced by more than 50% [30]. Under identical conditions, lpten 2 cells moved with an average velocity, average directional persistence, average chemotactic index (CI) and percent cells with a positive CI, statistically indistinguishable, using the student T-test (data not shown), from that of parental Ax2 cells and complemented lpten 2 /lpten oe cells ( Figure 3A). Computerreconstructed perimeter plots revealed similar directionality up a low cAMP concentration gradient ( Figure 3B, C, D), but lpten 2 cells were still defective in suppressing lateral pseudopod formation ( Figure 3C), as is evident when the perimeter tracks of mutant cells are compared to those of Ax2 ( Figure 3B) and lpten 2 /lpten oe cells ( Figure 3D). This was demonstrated in the analysis of the frequency of lateral pseudopod formation ( Figure 3E). 3D reconstructions revealed that the majority of the excess lateral pseudopods in cAMP gradients were formed by lpten 2 cells off the substratum (data not shown), as was evident in buffer in the absence of cAMP (Figure 2A, F). Lateral pseudopods that contact the substratum force turns [78]. Hence, the formation of excess lateral pseudopods off the substratum did not result in a significant increase in turning, which may explain why there was no significant decrease in the CI.

A. Parameters during cAMP chemotaxis.
Generating a ptenA 2 /lpten oe strain The similarities between the strong behavioral defects of the ptenA 2 mutant and the weaker defects of the lpten 2 mutant, suggested that the two homologs may play overlapping roles in basic cell motility and chemotaxis. We therefore tested whether overexpressing lpten in the ptenA 2 mutant would partially alleviate, or even rescue, the severe defects exhibited by ptenA 2 cells. The ptenA 2 mutant was transformed with an expression plasmid in which the cloned lpten coding region was placed under regulation of the strong actin15 promoter and fused in frame at its 39 end with red fluorescent protein (rfp) [79] ( Figure 4A). Expression of the entire 3.7 Kb lpten 2 rfp mRNA was verified using RT-PCR with primers P8 and P9 ( Figure 4A, Table S1). Upon achieving chemotactic responsiveness, aggregation-competent cells of the transformed line ptenA 2 /lpten oe expressed approximately 10 times as much lpten mRNA as the untransformed ptenA 2 mutant ( Figure 4B, inserted box). Overexpression of lpten rescued the developmental defects of the ptenA 2 mutant, resulting in normal aggregation (data not shown) and the formation of normal fruiting bodies (compare Figures 4C, D and E).   Overexpression of lpten rescues the basic behavioral and chemotactic defects of ptenA 2 cells As previously demonstrated [29,30], cells of the ptenA 2 mutant originally generated by Iijima and Devreotes [29], did not undergo morphogenesis on filter pads saturated with a K + -based buffer. Furthermore, when incubated on pads saturated with K + -based buffer to attain chemotactic competence and then assessed for basic motile behavior on the glass surface of a chamber perfused with K + -based buffer lacking cAMP, these cells crawled at less than half the average velocity of control cells and with less than half the directional persistence. These same characteristics were observed in the ptenA 2 mutant used here, which was generated by Hoeller and Kay [32] ( Figure 5A). The abnormal behavior of ptenA 2 cells was obvious, when computer-reconstructed perimeter tracks of control and ptenA 2 cells translocating in buffer were compared ( Figure 5B and C). The ptenA 2 cells translocating in buffer also formed lateral pseudopods at frequencies close to twice that of control cells ( Figure 5E), as was the case for the lpten 2 strains ( Figure 2G). Overexpression of lpten in ptenA 2 /lpten oe cells rescued every motility defect in the basic behavior of cells

Ax2
Cell no. migrating in buffer in the absence of chemoattractant ( Figure 5A), resulting in normal perimeter tracks ( Figure 5D) and restored suppression of lateral pseudopod formation ( Figure 5E). And, as previously demonstrated, ptenA 2 cells exhibited the same motility defects in a low cAMP concentration gradient, as they did in buffer alone, as well as a dramatic decrease in chemotactic responsiveness [29,30] (Figure 5F, H, J). Overexpression of lpten in ptenA 2 /lpten oe cells rescued every motility and chemotaxis defect ( Figure 5F), resulting in directed motility tracks up a gradient ( Figure 5I) and restored suppression of lateral pseudopod formation ( Figure 5J).
Overexpression of lpten rescues the ptenA 2 defect in natural aggregation Finally, as Wessels et al. [30] demonstrated, ptenA 2 cells [32] are defective in natural aggregation. In a natural aggregation territory, parental Ax2 cells moved in a highly directed ( Figure 6A) and cyclic fashion towards aggregation centers, increasing velocity in the front of each relayed, outwardly moving, non-dissipating wave of cAMP (velocity plots for two representative neighboring cells in the lower portion of Figure 6A). This resulted in centroid tracks that were directed at the source of chemotactic waves. Cells decreased velocity in the back of each wave, then reassessed directionality at the onset of the front of each wave, adjusting for deviations in direction during the translocation phase [48,51]. To assess aggregation centers in ptenA 2 cell populations, which do not complete normal aggregation, we retrospectively identified the point in each putative aggregation territory to which cells made net directional progress. Although the ptenA 2 cells on average did make net progress towards the interpreted aggregation centers ( Figure 6B, upper portion) and exhibited cyclic behavior (Figure 6B, lower portion), their centroid tracks were stunted and far less directional (i.e., less oriented on average in the direction of the interpreted aggregation center) ( Figure 6B, upper portion) and cycling was more erratic ( Figure 6B, lower portion). ptenA 2 cells tended to undergo far more directional changes than control cells, resulting from sharp turns away from the interpreted aggregation center. Overexpression of lpten in ptenA 2 /lpten oe cells restored normal behavior in a natural aggregation territory ( Figure 6C). Cells surged in the front of each wave ( Figure 6C, lower portion) and moved in a relatively directed fashion, with far fewer sharp turns, towards the aggregation center ( Figure 6C, upper portion), in a manner similar to that of parental Ax2 cells ( Figure 6A, upper portion).
Pulsing ptenA 2 cells up-regulates lpten and rescues chemotaxis in high, but not low, cAMP concentration gradients In performing this study, one apparent controversy had to be resolved. In three previous studies of ptenA 2 cell behavior [29,30,32], similar defects in velocity were described, but there was a lack of consensus on the capacity of mutant cells to assess a spatial gradient of cAMP generated in vitro. In all three studies, data were provided for cells that were induced to acquire chemotactic competence by a similar method of pulsing with cAMP [80,81] (Table 1), rather than incubating them on pads saturated with buffer, as performed for the experiments reported here (Table 1). All three studies employed buffers in the in vitro chemotaxis assays ( Table 1) that contained concentrations of cations that facilitated cAMP chemotaxis [48,51,59,77]. However, the studies differed in the concentration range of the cAMP gradients used. Iijima and Devreotes [29] analyzed responsiveness in a gradient of cAMP generated by releasing 1mM cAMP from a C.    [77].
c. When the source of the cAMP gradient was 1 mM, it generated a ''low cAMP concentration gradient'', in the concentration range estimated for the natural wave [44], and when it was 10 mM, it generated a high cAMP concentration gradient 10 times that estimated for the natural. doi:10.1371/journal.pone.0108495.t001 micropipette (low cAMP concentration gradient). They observed that ptenA 2 cells exhibited reduced velocity and a loss of chemotactic orientation (Table 1). Wessels et al. [30] analyzed chemotactic responsiveness in a gradient generated in a chamber [61,63], in which the source well was filled with 1 mM cAMP (low cAMP concentration gradient). They observed a 40% reduction in velocity and a 60% reduction in the chemotactic index ( Table 1). The gradients in these two studies were estimated, using the methods of Postma and Van Haastert [82], to be approximately 0.5 and 0.4 nM per mm and in the concentration range of the gradient estimated for the front of a natural wave in an aggregation territory (i.e., 1 mM at the peak of each wave and less than 0.01 mM at the trough) [44]. In contrast, Hoeller and Kay [32] analyzed responsiveness in a cAMP gradient generated by releasing 10 mM cAMP from a micropipette. The gradient generated was estimated to be approximately 5 nM per mm, steeper and in a concentration range 10 times higher than gradients tested by Iijima and Deveotes [29] and Wessels et al., [30]. Hoeller and Kay [32] (Table S1), prior to cAMP pulsing (1 hr), after cAMP pulsing for six hours (6 hr) and after cAMP pulsing with buffer for six hours (6 h efficiency between Ax2 and ptenA 2 cells, but did report a defect in velocity. These results suggested to us the possibility that pulsing ptenA 2 cells with cAMP might up-regulate expression of lpten to a level that provides them with the capacity to assess the gradient in the high cAMP concentration range, but not the gradient in the low cAMP concentration range. To explore this hypothesis, we analyzed the behavior of cAMP-pulsed parental cells and cells of the ptenA 2 strain of Hoeller and Kay [32], in gradients generated in a gradient chamber [58,61,63]. Behavior was analyzed in low cAMP concentration gradient, in which the source well contained 1 mM cAMP (Figures 7A and B) and in a high cAMP concentration gradient, in which the source well contained 10 mM cAMP ( Figures 7C, D).
In the low cAMP concentration gradient, cAMP-pulsed parental Ax2 cells underwent chemotaxis with high velocities and high chemotactic indices ( Figure 7A). And as previously reported [29,30] and shown here in Figure 5, both the instantaneous velocity and the CI of ptenA 2 cells were dramatically reduced ( Figure 7B). In a high concentration gradient, cAMPpulsed parental Ax2 cells underwent chemotaxis, but velocity was reduced by 25% and chemotactic efficiency (C.I.) by 35% ( Figure 3C), reductions similar to those previously reported by us using the same conditions thirty years ago [63]. However, in a high cAMP concentration gradient, ptenA 2 cells moved with a chemotactic index similar to that of Ax2 cell in a low concentration gradient, with slightly reduced velocity, as previously reported by Hoeller and Kay [32].
To test whether cAMP pulsing caused an increase in lpten expression in ptenA 2 cells, the level of the lpten transcript was compared between Ax2 and ptenA 2 cells by RT-PCR prior to pulsing (1 h), after six hours of pulsing with cAMP and after 6 hours of pulsing with buffer alone. Both in Ax2 cells ( Figure 7E) and ptenA 2 cells ( Figure 7F), cAMP pulsing up-regulated lpten expression at least 5 fold over that of the initial vegetative cell preparation (0 h). Pulsing ptenA 2 cells with buffer alone also upregulated lpten expression, but to a lesser degree than pulsing with cAMP ( Figure 7E, F). These results demonstrate that pulsing with cAMP up-regulates lpten expression, and, by correlation, may explain why cAMP-pulsed ptenA 2 cells can undergo chemotaxis in a high cAMP concentration gradient. The increased levels of expression of lpten 2 in cAMP-pulsed Ax2 and ptenA 2 cells were still several fold lower than the levels attained in strain ptenA 2 / lpten oe cells developed on pads (data not shown). This may explain why pulsing does not rescue the chemotaxis defect in a low cAMP concentration gradient.

Discussion
Mutations in the human PTEN gene are the most common of the tumor suppressor genes associated with human cancer [2,3,12,83]. Interestingly there are two additional PTEN homologs, TPTE and TPIPlocated on different chromosomes, and a secreted PTEN [12,13,[15][16][17][18], as well as a pseudogene of PTEN, PTENP1 [17]. However, there have been, to our knowledge, no reported studies to test whether any of the human PTEN homologs, when overexpressed, can rescue the behavioral defects caused by a human mutant PTEN cell, a question relevant to cancers that involve this mutation.
Here, we report for the first time that D. discoideum contains not only the human PTEN ortholog ptenA, as previously demonstrated [28,29], but also a second ortholog, lpten. Both PtenA and Lpten contain the two conserved domains of human PTEN, the dual-specificity phosphatase domain and PTEN-C2, the lipid-C2-binding domain. In addition, Lpten contains five LIM domains, which presumably play a role in protein-protein interactions [84,85]. Both ptenA and lpten are up-regulated in the period of the D.discoideum developmental program following the onset of starvation and preceding aggregation. Deletion of ptenA causes major defects in chemotaxis and development [29,30,32]. The ptenA 2 cells cannot undergo natural chemotaxis, aggregation or morphogenesis. However, deletion of lpten does not block aggregation or development, and does not decrease the efficiency of chemotaxis in a gradient of cAMP generated in vitro in the concentration range of the cAMP gradient in the front of the naturally relayed cAMP wave. Deletion of lpten does, however, affect the suppression of lateral pseudopod formation, which is also the case for the ptenA 2 mutant. The mutant phenotype of lpten 2 , therefore, exhibits a weak phenocopy of ptenA 2 .
Here we have considered the possibility that there may exist at least parallel functions, or partial redundancy of Pten and Lpten. This has led us to test whether overexpressing lpten in a ptenA 2 mutant might rescue the defects of the ptenA 2 mutant. We therefore placed the coding region of lpten under the control of the strong actin 15 promoter in a plasmid and introduced it into the ptenA 2 mutant to create ptenA 2 /lpten oe . Aggregation-competent cells of ptenA 2 /lpten oe expressed the lpten transcript at over 10 times the level observed in wild type or ptenA 2 cells, and close to two orders of magnitude higher than in vegetative cells. We found that overexpression rescued every developmental, cell motility and chemotaxis defect exhibited by mutant ptenA 2 cells [30]. The rescued ptenA 2 defects included the following: 1) a prolonged preaggregative period; 2) abnormal or no aggregation in submerged cultures or on a filter pad substrate; 3) lack of a multicellular developmental program, including lack of fruiting body formation; 4) decreased velocity during basic motile behavior in buffer alone, or during chemotaxis in a shallow gradient of cAMP generated in the concentration range estimated for the front of a natural wave [44] 5) an increased frequency of lateral pseudopod formation and turning; 6) a dramatic decrease in chemotactic efficiency in a low cAMP concentration gradient, in the estimated range for the front of the natural wave; and 7) a dramatic decrease in orientation in the front of relayed cAMP waves in a natural aggregation territory. The complete rescue of the ptenA 2 mutant phenotype by overexpression of the homolog lpten, raises the possibility that overexpressing a human PTEN homolog might suppress the behavioral defects of a mutant PTEN thus suppressing tumorigenesis.

Supporting Information
Table S1 Primers used in this study. (PDF)