ADP-ribosylation factor machinery mediates endocytosis in plant cells

Abstract

Endocytosis is crucial for various cellular functions and development of multicellular organisms. In mammals and yeast, ADP-ribosylation factor (ARF) GTPases, key components of vesicle formation, and their regulators ARF-guanine nucleotide exchange factors (GEFs) and ARF-GTPase-activating protein (GAPs) mediate endocytosis. A similar role has not been established in plants, mainly because of the lack of the canonical ARF and ARF-GEF components that are involved in endocytosis in other eukaryotes. In this study, we revealed a regulatory mechanism of endocytosis in plants based on ARF GTPase activity. We identified that ARF-GEF GNOM and ARF-GAP VASCULAR NETWORK DEFECTIVE 3 (VAN3), both of which are involved in polar auxin transport-dependent morphogenesis, localize at the plasma membranes as well as in intracellular structures. Variable angle epifluorescence microscopy revealed that GNOM and VAN3 localize to partially overlapping discrete foci at the plasma membranes that are regularly associated with the endocytic vesicle coat clathrin. Genetic studies revealed that GNOM and VAN3 activities are required for endocytosis and internalization of plasma membrane proteins, including PIN-FORMED auxin transporters. These findings identified ARF GTPase-based regulatory mechanisms for endocytosis in plants. GNOM and VAN3 previously were proposed to function solely at the recycling endosomes and trans-Golgi networks, respectively. Therefore our findings uncovered an additional cellular function of these prominent developmental regulators.

The ADP-ribosylation factor (ARF) GTPase machinery is the regulatory system that controls endocytosis in metazoans and fungi. The ARF GTPases govern vesicle budding, and their activity is regulated spatiotemporally by ARF-guanine nucleotide exchange factors (GEFs) and ARF-GTPase-activating protein (GAPs). ARF-GEFs facilitate the exchange of bound GDP for GTP for ARF GTPase, whereas ARF-GAPs stimulate the GTP hydrolysis activity of ARF GTPase to counteract the ARF-GEF activity that, in turn, facilitates the ARF GTPase cycle (1). In metazoans and fungi, ARF-GEFs and ARF-GAPs consist of various groups, and specific ARF-GEFs and ARF-GAPs are used to control the ARF GTPase at the plasma membrane (PM) (2). For instance, large-sized Golgi-specific Brefeldin A resistance factor (GBF)-type ARF-GEFs are involved in endoplasmic reticulum (ER)-to-Golgi trafficking, and small- and medium-sized ARF-GEFs that contain the pleckstrin homology domain are involved in endocytosis (Fig. S1 A and B) (1, 3, 4). However, in the plant genomes, the small- and medium-sized ARF-GEFs are absent (Fig. S1 A and B) (3). These differences suggested that the regulatory mechanism of endocytosis in plants has evolved differently from that in other eukaryotes.

In general, plants share components of the ARF GTPase machinery. The functional characterization of several Arabidopsis thaliana ARF-GEFs and ARF-GAPs has revealed their crucial roles in plant development. The ARF-GEF GNOM that belongs to the large-sized GBF type of ARF-GEFs is essential for plant development, particularly for processes such as embryo and seedling patterning (5, 6) that depend on the transport of the plant hormone auxin, and has been shown to regulate the recycling of PIN-FORMED (PIN) auxin transporters from the endosomes to the PM (7, 8). Similarly, the ARF-GAP VASCULAR NETWORK DEFECTIVE 3 (VAN3) controls auxin transport-mediated processes such as vascular tissue formation (9, 10), and its cellular role, although not entirely clear, has been mapped to the endosomal compartments (9, 11). In contrast to the extensive data in yeast and mammals (4, 12, 13), very little is known about the direct involvement of the ARF machinery in the endocytosis of plants (14–16). It has been suggested that the GNOM-like 1 (GNL1) ARF-GEF, the closest homolog of GNOM, regulates the internalization of the PIN2 auxin transporter (17), but its action remained unclear because the loss-of-function gnl1 mutants did not show any endocytosis defects per se (18). The absence of canonical endocytic ARF components in plant genomes as well as the lack of any data demonstrating the presence of ARF GEFs and ARF GAPs at the PM of plant cells raises a question about how endocytosis regulation is realized in plants.

Results and Discussion

van3 Mutants Phenocopy gnom Mutants.

To examine the functions of plant ARF GTPases and their regulators, we studied the developmental and cellular roles of ARF-GEF GNOM and ARF-GAP VAN3. VAN3/SCARFACE (SFC) and its three homologs, the VAN3-like (VAL) proteins, redundantly regulate the formation of vasculature, whereas val1val2val3 triple mutants did not show remarkable phenotypes (10). We further analyzed the development of plants lacking the function of VAN3 and related proteins. Besides the venation discontinuity (Fig. S1 C–F), sfc-9val1val2val3 mutants had fused lateral root primordia (33.1%, n = 124) and defective cotyledon formation (4.1%, n = 763) (Fig. 1 A–C). The whole range of developmental defects were strongly reminiscent of those found in weak gnom mutant alleles that are defective in GNOM ARF-GEF (19). The phenotypic similarity and the common biochemical function in regulating ARF GTPases suggest a functional relation between VAN3 and GNOM in the same process.

Fig. 1.

Similarity of gnom and van3 phenotypes. A–C show seedling phenotypes of sfc-9val1val2val3 mutants. (A) Wild type (Col) (Left), sfc-9val1val2val3 quadruple (quad) (Center), and sfc-9 mutant (Right). Insets show magnified views of gnom^R5 (gn^R5) and sfc-9val1val2val3 quadruple mutants (quad). (B and C) Lateral root primordia development in (B) wild type (Col) and (C) sfc-9val1val2val3 mutant (quad).

VAN3 ARF-GAP and GNOM ARF-GEF Are Localized at the PM.

To assess whether VAN3 and GNOM can act together in the cell, we examined their subcellular distribution. Previous reports placed VAN3 in the trans-Golgi network (TGN)/early endosome (16) in cultured cells and mature tissues of Arabidopsis (9, 11). In contrast, in developing organs, such as root meristems, embryos, and lateral root primordia, the functional VAN3-GFP, which was expressed in the van3 mutant and complemented its phenotype, localized preferentially at the PM, in addition to the minor intracellular dot-like signals (Fig. 2 A-C).

Fig. 2.

Localization of the ARF GTPase machinery at the PM. (A–C) Localization of the functional VAN3-GFP at the PM. (A) Heart-stage embryo. (B) Root cortex cells. (C) Lateral root primordium. (D–F) Confocal images of seedlings coexpressing (D) functional GNOM-GFP and (E) VAN3-mRFP, showing intracellular and PM localization and preferential PM labeling as well as minor dot-like structures, respectively. (F) Merged image of GNOM-GFP (green) and VAN3-mRFP (red). Inset shows nonoverlapping intracellular localization of GNOM and VAN3. (G–I) Confocal images of seedlings coexpressing GNOM-GFP and VAN3-mRFP in BFA-treated (40 μM, 1 h) root cells. (G) GNOM-GFP in BFA-treated cells revealing strong labeling of BFA compartments and the PM. (H) VAN3-mRFP in BFA-treated cells showing the PM signal. (I) Merged images of GNOM-GFP and VAN3-mRFP in BFA-treated root cells. Arrows and arrowheads mark the endosomal localization and the signal at the PM, respectively. [Scale bars: 10 μm (A), 12 μm (B), 20 μm (C), and 4 μm (D–I).]

Because GNOM functions at recycling endosomes (7), VAN3 and GNOM might act intracellularly at the same endosomes. To explore this possibility, we examined the localization of the functional GNOM-GFP and VAN3-monomeric red fluorescent protein (mRFP) expressed in the embryo-defective30-1 and sfc-9 mutants, respectively. However, simultaneous localization of VAN3-mRFP and GNOM-GFP proteins revealed no intracellular colocalization (Fig. 2 D–F). Unexpectedly, careful observation identified a weak, but regular, localization of the functional GNOM-GFP at the PM of different cell types (12.2 ± 3.3% of seedlings, n = 83) (Fig. 2D and Table S1). ARF-GEFs are known to dissociate rapidly from the membrane after their action (20, 21), and this rapid dissociation typically hampers the observation of ARF GEFs, including GNOM, at their place of action. To address further the presence and action of GNOM at the PM, we applied the ARF-GEF inhibitor brefeldin A (BFA) that is known to stabilize BFA-sensitive ARF-GEFs at the membranes where they act (22). In Arabidopsis, BFA treatment mediates GNOM accumulation at recycling endosomes and their aggregation with the TGN to form “BFA bodies” (23) where various PM proteins accumulate as a result of protein recycling inhibition (7, 24). Notably, after the BFA treatment, strong GNOM-GFP labeling was observed not only in BFA bodies but also at the PM (100 ± 0% of seedlings, n = 68) (Fig. 2G and Table S1), as can be seen in the previously published micrographs (7). Similarly, VAN3-mRFP was still present at the PMs after BFA treatment (Fig. 2 H–I). To compare the incidence of GNOM at the PM and in the BFA bodies and to exclude potential secondary effects, we examined the GNOM-GFP localization at different time points and concentrations of BFA treatment. To observe better the effect of BFA on the GNOM-GFP localization, we excited GNOM-GFP at low laser power that could not detect the clear PM-localized GNOM-GFP before the BFA treatment. Live-cell imaging of GNOM-GFP after BFA application revealed that GNOM-GFP associated strongly with the PM at the same time as or even before the formation of BFA bodies (Fig. S2A). Moreover, at low BFA concentrations, the BFA bodies were not visibly formed, but GNOM-GFP showed a strong decoration of the PM (Fig. S2B). Importantly, the engineered, BFA-resistant version of GNOM-MYC (7) was not stabilized at the PM after BFA treatment (Fig. S2 C and D and Table S1). These results confirm the localization of GNOM together with VAN3 at the PM. Moreover, the stabilization of GNOM at the PM after the BFA treatment implies that this localization has functional significance and that the ARF machinery acts at the PM. These results suggest that VAN3 and GNOM do not colocalize in the intracellular endomembranes but localize together at the PM and that they have a common function there.

GNOM ARF-GEF and VAN3 ARF-GAP Are Partially Colocalized and Recruited to Clathrin Foci at the Cell Surface.

To gain further information on the action of the GNOM ARF-GEF and VAN3 ARF-GAP at the PMs, we performed variable angle epifluorescence microscopy (VAEM) analysis that enables the selective visualization of cell surfaces (25, 26). GNOM-GFP and VAN3-mRFP were organized prominently into discrete foci at the cell periphery. Foci of GNOM-GFP and VAN3-mRFP partially overlapped, although some independent foci were present (Fig. 3 A–C). Approximately 40% of all VAN3-mRFP foci (n = 304) overlapped with the GNOM-GFP foci, whereas ≈40% of all of the GNOM-GFP foci (n = 309) overlapped with the VAN3-mRFP foci (Fig. 3J). To exclude the possibility that the colocalization was the result of random overlapping of the foci in the cell cortex of epidermal cells, we applied a previously used technique (27, 28) in which, as a negative control, the colocalization rate was counted by rotating the red channel image with respect to the green channel image by 180°. The average colocalization rate of GNOM and VAN3 was much higher than that of the rotated images, indicating that their colocalization was highly significant (Fig. 3J). Thus, GNOM and VAN3 partially colocalize, implying a functional relationship between VAN3 and GNOM and suggesting that GNOM and VAN3 have both overlapping and distinct functions at the PMs.

Fig. 3.

VAEM micrographs of VAN3 and GNOM. (A–C) Double labeling of GNOM-GFP (A) and VAN3-mRFP (B) and merged image (C). GNOM-GFP and VAN3-mRFP colocalized at foci on the PM (arrowheads). (D–F) Double labeling of GNOM-GFP (D) and CLC-mKO (E) and merged image (F). GNOM-GFP and CLC-mKO colocalized at foci on the PM (arrowheads). (G–I) Double labeling of VAN3-GFP (G) and CLC-mKO (H) and merged image (I). VAN3-GFP and CLC-mKO colocalized at foci on the PM (arrowheads). (J) Colocalization efficiency of GNOM-GFP and VAN3-mRFP. (Left) Percentage of GNOM-GFP foci colocalized with VAN3-mRFP. (Right) Percentage of VAN3-mRFP foci colocalized with GNOM-GFP. As a negative control, the percentage of random colocalization (red frame is turned for 180°) is given. (K) Colocalization efficiency of GNOM-GFP and CLC-mKO. Left and right graphs give the percentage of GNOM-GFP foci colocalized with CLC-mKO and CLC-mKO foci colocalized with GNOM-GFP, respectively. As a negative control, the percentage of random colocalization (red frame is turned for 180°) is given. (L) Colocalization efficiency of VAN3-GFP and CLC-mKO. Left and right graphs give the percentage of VAN3-GFP foci colocalized with CLC-mKO and CLC-mKO foci colocalized VAN3-GFP, respectively. As a negative control, the percentage of random colocalization (red frame is turned for 180°) is given. (Scale bars in A–I: 1 μm.)

VAEM may detect signals that are in close proximity to the PM as well as signals at the PM itself. To confirm that GNOM and VAN3 foci localized to PMs, we compared the localizations of GNOM and VAN3 with clathrin, which is localized to the PM to mediate endocytosis (29). We used the lines overexpressing the fluorescently labeled clathrin light chain (CLC), which is used as a marker of endocytosis site at the PMs (26, 29–31). Double labeling of GNOM-GFP and VAN3-GFP with monomeric Kusabira Orange (mKO)-tagged CLC revealed that GNOM-GFP and VAN3-GFP foci were observed in the same focal plane as CLC-mKO and that some GNOM-GFP and VAN3-GFP foci colocalized with mKO-CLC (Fig. 3 D–I, K, and L). These findings indicate that GNOM and VAN3 are located on the PMs, where some of them associated with the clathrin foci, further suggesting that these ARF regulators function in endocytosis.

GNOM and VAN3 Are Required for Endocytosis.

Subcellular localization analysis of VAN3 and GNOM suggested that they play a role in endocytosis. To test this hypothesis, we analyzed the uptake of the endocytic tracer N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl)pyridinium dibromide (FM4-64) (32–35) in van3 and gnom mutants. Pulse-labeling and time-lapse monitoring of the intracellular fluorescence revealed that FM4-64 uptake was clearly lower in van3-1 and van7 (weak alleles of gnom) roots than in the wild type (Fig. 4 A–C and G). A similar reduction of the FM4-64 uptake was also detected in other van3 and gnom alleles, including gnom^R5, sfc-9, and sfc-9val1val2val3 (Fig. S3 A–H).

Fig. 4.

VAN3 and GNOM involvement in endocytosis. (A–C) Reduced uptake of the endocytic tracer FM4-64 (5 min incubation) in van3-1 (B) and weak gnom (van7) (C) mutants compared with the wild-type Ler (A). (D–F) Inhibition of the internalization of PIN1-GFP in response to BFA (30 min, 50 μM CHX followed by 60 min of 50 μM CHX/40 μM BFA) in van3 (sfc-9) (E) and gnom^R5 (F) mutant roots compared with the wild-type Col-0 (D). (G) Quantification of FM4-64 uptake in untreated van3 and van7 mutants. Quotients of the fluorescence signal intensity between the inside of the cells (excluding the PM) and the PMs were calculated and given in the ordinate. Note the inhibited uptake of FM4-64 in van3 and weak gnom (van7) root cells. (H and I) Inhibited internalization of FM4-64 in BFA-treated roots expressing the BFA-sensitive GNOM (GN-sens) (H) compared with the BFA-resistant GNOM (GN-Res) (I). The comparison was done in the background of gnl1 mutants expressing the BFA-sensitive GNL1. Seedlings were pretreated with 40 μM BFA for 30 min and were pulse labeled with FM4-64. Time points after the FM4-64 pulse are indicated. (Scale bars in A–I: 4 μm.)

Next, we tested the involvement of VAN3 and GNOM in the internalization of proteins from the PM. To visualize the protein internalization, we applied BFA and monitored the intracellular accumulation of the endocytosed PIN1 auxin transporter (8). BFA prominently inhibits the recycling pathway rather than the endocytosis pathway, so that the ratio between endocytosis and recycling is perturbed and the endocytosis material accumulates intracellularly (7). To exclude any influence of de novo protein synthesis, we coincubated samples with the protein synthesis inhibitor cycloheximide (CHX). In both the presence and absence of CHX, PIN1-GFP internalization and intracellular accumulation were strongly reduced in sfc-9 and gnom^R5 mutants when compared with wild-type roots (Fig. 4 D–F and Fig. S4A). Similarly, anti-PIN1 antibody staining revealed a diminished PIN1 internalization in roots of van7 and sfc-9val1val2val3 roots (Figs. S3 I and J and S4B). These results suggested that the endocytosis-to-recycling ratio was smaller in the van3 and gnom mutants than in the wild type.

Because the BFA compartments in wild-type cells contain endocytosis material from the PM (29), endocytosis appears to be largely operational under the inhibition of the GNOM-dependent trafficking by BFA (Fig. 4D), suggesting that BFA-resistant ARF-GEF are involved in endocytosis. Because GNL1, a BFA-resistant ARF-GEF, had been shown to act redundantly with GNOM in the ER-to-Golgi trafficking (36), we examined whether BFA-resistant GNL1 also might function redundantly with GNOM in endocytosis. The FM4-64 uptake analysis did not reveal any endocytosis defects in BFA-untreated gnl1 mutant roots (Fig. S3 K and L), consistent with previous observations (18); in contrast, the BFA treatment caused PIN1 internalization defects in the gnl1 mutants (Figs. S3 M and N and S4C) (17). Possible redundant function between BFA-sensitive GNOM and BFA-resistant GNL1 made it difficult to dissect their role in endocytosis after the BFA treatment. ARF-GEF sensitivity or resistance to BFA can be changed by introducing a mutation that affects the sensitivity to BFA but not the overall protein function (37), providing a means to manipulate specifically the function of a selected ARF GEF (36). To clarify this apparent discrepancy and assess a GNOM function in endocytosis, we used the engineered versions of GNOM and GNL1 with changed BFA sensitivities (7, 36). We used the background of the BFA-sensitive version of GNL1 to inhibit the GNL1 function by BFA treatment and manipulate only the GNOM function. In the background of the BFA-sensitive GNL1, the BFA-sensitive GNOM reduced the BFA-induced PIN1 internalization in contrast to the BFA-resistant GNOM (Figs. S3 O and P and S4D). Similarly, the FM4-64 fluorescence from the PM disappeared more slowly in the presence of BFA when GNOM was BFA sensitive than when it was BFA resistant (Fig. 4 H and I and Fig. S3 Q and R).

Together with the findings on FM4-64 uptake defects in untreated gnom mutants but not gnl1 mutants (Fig. 4C and Fig. S3L) (18), these results show that it is primarily GNOM that is required for endocytosis, both in the presence and absence of the GNL1 function. On the other hand, in the absence of the GNOM function (BFA treatment of the wild type), GNL1, which is resistant to BFA, partially takes over the function of GNOM in endocytosis, because there are additional internalization defects in the BFA-treated gnl1 mutants (17) resulting from the inhibition of both GNOM and GNL1 function at the PM. Thus, both GNOM and GNL1 are required for endocytosis, with GNOM being the major factor in this process.

Conclusions

Here we identified an additional function of ARF-GEF GNOM and ARF-GAP VAN3 in endocytosis at the PMs, besides their function at intracellular organelles (5, 9). The phenotypic similarity between the multiple van3 mutants and the weak allele of the gnom mutants prompted us to examine the functional relationship between GNOM and VAN3. Surprisingly, these regulators did not colocalize in the intracellular compartments but localized together to the PM. VAEM analysis confirmed that a large portion of GNOM-GFP and VAN3-mRFP foci were located in the same place at the PM and colocalized regularly with clathrin foci. Importantly, genetic inhibition of the GNOM or VAN3 functions leads to defects in endocytosis. So far, the focus had been put on the intracellular localization and function of GNOM and VAN3, but their role in endocytosis had not been proposed (7, 9). Nonetheless, our results reveal both the presence of these regulators at the PM and the requirement of their function for endocytosis. These findings, in particular, in the case of the GNOM function, are unexpected and, to some extent, contradict previous interpretations (7) but are not inconsistent with previously reported data. Linear sucrose-density gradient centrifugation detected the GNOM protein in the PM fractions as well as in other membrane fractions, although the GNOM action at the PMs had never been mentioned (6). Furthermore, re-evaluation of previous localization studies also revealed a faint, but consistent, localization of GNOM at the PM (7, 38). Finally, expression of GTP- and GDP-fixed forms of ARF1 inhibits endocytic FM4-64 uptake (39), also suggesting the involvement of the ARF GTPase machinery in endocytosis. Our work clarifies the apparent previous inconsistencies, provides additional insights into the role of GNOM ARF-GEF and VAN3 ARF-GAP, and redefines the cellular function of these two important developmental regulators.

Our findings also have important implications for the evolution of the membrane-trafficking machineries in plants. We show that the ancestral GBF-type ARF-GEFs are used for regulating endocytosis in Arabidopsis, although in fungi and metazoans they are specifically involved in the ER-to-Golgi traffic (2). Because plants lack the class III ARFs that act at the PM in mammals (2, 40), it would be interesting to analyze which ARF proteins are used to regulate endocytosis in plant cells. Our work suggests that the recruitment of ARF-GEF regulators for various ARF GTPase functions also differs from that in metazoans and fungi. These findings point out important differences in the mechanism of ARF-dependent endocytosis between plants and other eukaryotes and also open avenues for investigating the independent coevolution of ARF-based vesicle trafficking machineries in different eukaryotes.

Materials and Methods

Plant Material and Growth Conditions.

Arabidopsis thaliana (L.) Heynh. seedlings harboring the following transgenes and mutants have been described previously: pPIN1::PIN1-GFP (39, 41), pGNOM::GNOM-GFP (7), pVAN3::VAN3-GFP (11), BFA-resistant GNOM–MYC (7), BFA-sensitive GNL1-MYC (17, 36), BFA-sensitive GNL1-YFP (17, 36), van3-1 (9), sfc-9 (10), van7 (42), and gnom^R5 (19). gnom^R5 expressing PIN1-GFP was back-crossed four times to the Columbia (Col) ecotype. All the pVAN3::VAN3-xFP and pGNOM::GNOM-GFP constructs were expressed in their own mutants and tested for functionality. To analyze seedling phenotypes, seeds were sown on 0.8% agar-solidified 0.5× Murashige and Skoog medium supplemented with 1% sucrose (pH 5.9) at 18 °C and a 16-h/8-h light/dark cycle, unless otherwise indicated. Detailed information is given in SI Materials and Methods.

Plasmid Construction.

The functional pVAN3::VAN3-mRFP was made as described (11). The 35S::CLCmKO was constructed by inserting the mKO-tagged (43) CLC fragment into pBI101.

Histological Analysis.

Ten-day-old seedlings were used for observation of the veins and lateral root primodia. Seedlings were fixed in a 9:1 mixture of ethanol and acetic acid solution for 1 h at room temperature. Fixed samples were cleared by chloral hydrate solution (8 g chloral hydrate: 1 mL glycerol: 2 mL distilled water) and then observed directly by a BX 51 light microscope (Olympus).

Antibody Staining and Confocal Laser Scanning Microscopy.

Whole-mount immunolocalization on Arabidopsis roots was done by an Intavis in situ Pro automated system as described previously (44). Antibodies were diluted as follows: rabbit anti-PIN1 antibody, 1:2,000 (45); Cy3-conjugated secondary anti-rabbit antibody, 1:600 (Sigma). For FM4-64 labeling of Arabidopsis seedling roots, 4- to 5-d-old seedlings were pulse labeled in liquid medium containing 2 μM FM4-64 for 5 min at 4 °C. Immunofluorescence and live-cell microscopy were done with an Olympus FV1000.

VAEM Observation.

Root epidermal cells of 5-d-old seedlings were observed under a fluorescence microscope (Nikon Eclipse TE2000-E and a CFI Apo TIRF 100× H/1.49 numerical aperture objective) equipped with a Nikon TIRF2 system. GFP and mKO/mRFP were excited with a 488-nm and 561-nm laser, respectively. The fluorescence emission spectra were separated with a 565LP dichroic mirror and filtered through either a 515/30 (GFP) or 580LP (mKO and mRFP) filter with a Dual View filter system (Photometrics). All images were acquired with an Andor iXonEM EMCCD camera.

Quantification of FM4-64 Uptake.

The mean fluorescence intensity of vesicles inside the cells, excluding the PM, was measured with ImageJ (http://rsbweb.nih.gov/ij/). The quotients between the inside and the PM value were calculated for 10–15 cells from at least three different roots. The values of three independent experiments were averaged. The value of each phenotype was standardized to the corresponding wild-type control to evaluate the percentage of endocytosis inhibition compared with a wild-type condition.