Genetic analyses of the Arabidopsis 26S proteasome regulatory particle reveal its importance during light stress and a specific role for the N-Terminus of RPT2 in development.

Abstract

The 26S proteasome subunit RPT2 is a component of the hexameric ring of AAA-ATPases that forms the base of the 19S regulatory particle (RP). This subunit has specific roles in the yeast and mammalian proteasomes by helping promote assembly of the RP with the 20S core protease (CP) and gate the CP to prevent indiscriminate degradation of cytosolic and nuclear proteins. In plants, this subunit plays an important role in diverse processes that include shoot and root apical meristem maintenance, cell size regulation, trichome branching, and stress responses. Recently, we reported that mutants in RPT2 and several other RP subunits have reduced histone levels, suggesting that at least some of the pleiotropic phenotypes observed in these plants result from aberrant nucleosome assembly. Here, we expand our genetic analysis of RPT2 in Arabidopsis to shed additional light on the roles of the N- and C-terminal ends. We also present data showing that plants bearing mutations in RP subunit genes have their seedling phenotypes exacerbated by prolonged light exposure.

Introduction

Protein breakdown plays an important role in all cellular organisms in the selective control of key regulatory proteins and in the removal of aberrant polypeptides and normal proteins deemed no longer necessary. This turnover is particularly pertinent in plants, which as sessile organisms are more susceptible to environmentally induced perturbations in protein folding.¹ Consequently, plant cells continually scan their protein complement for appropriate proteolytic targets to help restore protein homeostasis and to trigger the onset of key developmental checkpoints. Most often the ubiquitin/26S proteasome system (UPS) is used to direct selective turnover. It employs a highly polymorphic ATP-dependent conjugation cascade that attaches ubiquitin to one or more accessible lysine residues, either within the target substrate or on previously appended ubiquitin moieties.² The resulting polyubiquitylated proteins are then recognized and degraded by the multicatalytic 26S proteasome complex.³

The 26S proteasome is composed of two functionally distinct complexes, the 20S core protease (CP) capped on one or both ends with the 19S regulatory particle (RP). The CP is a broad-spectrum protease assembled from four stacked heteroheptameric rings containing either seven α-subunits or seven β-subunits (PAA-PAG and PBA-PBG in Arabidopsis) in a C2 symmetric α_1–7β_1–7β_1–7α_1–7 configuration. Upon assembly, a central chamber is created that houses six peptidase catalytic sites provided by the β₁ (PBA), β₂ (PBB), and β₅ (PBE) subunits,⁴ with access to this chamber being restricted by two narrow axial pores generated by the peripheral α-subunit rings. This architecture ensures that proteolysis is restricted only to those polypeptides that are deliberately unfolded and imported into the CP. The 20 or more subunit RP docks onto the CP by virtue of specific contacts with the α-subunit rings,^5-7 imparting both ATP-dependence and substrate specificity to the CP. The RP can be further subdivided into base and lid subcomplexes.³ The RP base consists of a heterohexameric ring of AAA-ATPases (RPT1–6) that directly contacts the α-subunit ring of the CP,⁷ along with three non-ATPase subunits, RPN1, RPN2 and RPN10. ATP binding to the RPT ring promotes association of the RP with the CP, whereas ATP hydrolysis drives substrate unfolding and subsequent import into the CP.⁸ The RP lid includes at least 10 additional subunits (RPN3, 5–9, 11–13 and 15). With the exceptions of RPN10, 11 and 13, the function(s) of most other RPN subunits remain largely obscure.³

Despite the 6-fold pseudosymmetry of the RPT ring, previous biochemical studies have indicated a special role for the RPT2 subunit in stabilizing the asymmetric interface between the RP and CP.^5,9 In particular, studies on the mammalian¹⁰ and yeast⁷ complexes revealed that RPT2, along with RPT3 and RPT5, help dock the RPT ring to the CP via the insertion of a C-terminal HbYX sequence (in which Hb is a hydrophobic residue, Y is tyrosine, and X is any amino acid) into a cleft at the interface between two α-ring subunits, which in the case of RPT2 are α₃ and α₄. Recent structural studies from yeast revealed that the RPT ring then forms a “spiral staircase” structure, with RPT2 occupying the lowest position, thus giving it the closest contact with the CP α-ring.¹¹

Our recent studies in Plant Cell described in-depth genetic analysis of the orthologous RPT2 protein in the flowering plant Arabidopsis thaliana,¹² which provided further support that this subunit has one or more unique activities within the RPT ring. Here, we present additional deletion analyses of the RPT2a isoform showing that N-terminus is more important than the exposed C-terminal HbYX motif for the phenotypic functions of RPT2. We also provide morphological data demonstrating that the phenotypes of various RP subunit mutants are at least partially rescued when plants are grown in short-day photoperiods, suggesting that compromised 26S proteasome activity becomes especially acute under high light conditions.

Results

The C-Terminal End of RPT2 is not Critical for Function

Unlike yeast in which all proteasome subunits are encoded by single genes, most RP and CP subunits in plants such as Arabidopsis are synthesized from two paralogous genes that often encode proteins with high amino acid sequence identity.¹³ For example, the RPT2 subunit is encoded by two loci (RPT2a and RPT2b) that express proteins with 99% amino acid similarity, both of which are incorporated into the 26S particle.¹⁴ Consistent with their high amino acid identity, but in contrast to previous reports,¹⁵ our genetic analyses, coupled with complementation studies using full-length coding sequences, revealed that the RPT2a and RPT2b isoforms are functionally equivalent.¹² Despite this redundancy, null rpt2a mutants display a range of aberrant phenotypes not seen with null rpt2b mutants, including problems with respect to gametophyte and sporophyte development, shoot and root apical meristem activity, cell elongation, DNA endoreduplication, trichome branching, responses to stress, and nucleosome assembly. This unique impact is likely caused by differences in expression level between the two genes, and by the fact that the RPT2a promoter alone responds to challenged proteasome capacity by increasing expression as part of the ‘proteasome stress’ regulon.¹² By comparing the phenotypes of rpt2a mutants with mutants affecting other RP subunits (e.g., rpn1a, rpn5a, rpn10, and rpn12a), we also confirmed that RPT2 has a unique function in plants, either as a component of the 26S proteasome, or by virtue of additional non-proteasomal activities.

Arabidopsis seedlings with diminished RPT2 levels also have difficulties assembling 26S particles from the RP and CP. As with other RP mutants, such as rpn5 and rpn7 in yeast¹⁶ or rpn5a and rpn12a in Arabidopsis,¹⁷ null rpt2a mutants have substantially reduced amounts of the complete 26S proteasome complex, with concomitantly increased amounts of free RP and CP. By contrast, null rpt2b mutants display normal proteasome assembly, presumably due to the increased expression of RPT2a maintaining a sufficient pool of RPT2 for normal 26S complex assembly. Prior C-terminal deletion analyses (RPT2a-ΔC3) suggested that part of the stabilizing influence of RPT2a on RP-CP assembly was manifested through its HbYX motif.¹² Despite these connections, two recent studies reported that the HbYX residues of mammalian RPT2 are in fact dispensable for assembly of the 26S proteasome,^18,19 leaving the importance of the HbYX motif in doubt, at least for RPT2.

To further study the role of the RPT2 C-terminus, we attempted to complement the rpt2a-2 mutant with an RPT2a protein missing the last eight amino acids and with RPT2 variants in which their C-termini were capped with additional sequence provided by the 14 residue T7 or 12 residue Myc epitopes. As shown in Figures 1 and 2, the RPT2-ΔC8, RPT2a-T7 and RPT2b-Myc proteins are functional, as judged by their ability to rescue most, if not all, of the phenotypic defects in rpt2a-2 plants. Evidence for expression of the transgenic proteins was provided by an increase in the total RPT2 protein detected immunologically as compared with that in rpt2a-2 seedlings (Figs. 1E and 2E). Detection of the intact RPT2a-T7 and RPT2b-Myc proteins with anti-T7 or anti-Myc antibodies, respectively, also confirmed that the tags remained attached to the RPT2 polypeptides in planta (Fig. 2E). Phenotypically, multiple independently transformed lines expressing the three RPT2a variants rescued the short stature of mature plants, reversed the lanceolate shape of rosette leaves, restored normal root growth, and reduced the excessive trichome branching characteristic of the rpt2a-2 background (Figs. 1A-D and 2A-D). For the tagged versions of RPT2, phenotypic rescue occurred regardless of which isoform was expressed, or which promoter was used. Expression of the RPT2-ΔC8, RPT2a-T7 and RPT2b-Myc proteins also reversed the proteasome stress seen in rpt2a-2 plants, as evidenced by the reduced accumulation of RPN1, RPN12a and PBA1 proteins down to wild-type levels, as compared with the elevated levels in rpt2a-2 plants (Figs. 1E and 2E). Collectively, our data indicate that an exposed C-terminal HbYX motif in RPT2 is not essential for either RPT2 activity or RP function. It is possible that sufficient RP-CP assembly takes place for the RPT ring to be tethered only by the HbYX motifs in RPT3 and RPT5, or that the phenotypes seen for rpt2a mutants reflect non-proteasomal functions of RPT2 and the RPT ring.

Figure 1. C-terminal deletion of Arabidopsis RPT2a is fully functional. A transgene expressing RPT2a-ΔC8 from the native RPT2a promoter was generated by overlapping PCR and cloned into the complementation vector pCAMBIA3301 as previously described.¹² Heterozygous rpt2a-2 mutant plants were transformed by Agrobacterium tumefaciens-mediated floral dip, and positive transformants were allowed to self. rpt2a-2 plants homozygous for both the rpt2a-2 T-DNA and the RPT2a transgenes were identified by PCR genotyping and BASTA resistance, respectively. (A) Phenotypic rescue of inflorescence development. Wild-type (WT), rpt2a-2 and three independent homozygous rpt2a-2 rescue lines expressing RPT2a-ΔC8 were grown in LD for 7 weeks. (B) Phenotypic rescue of abnormal leaf shape. Pictured are newly expanded leaves taken from plants grown in LD for 7 weeks. (C) Phenotypic rescue of primary root growth. Each bar represents the average root length (± SD) of 20 plants grown in LD for 14 d. (D) Phenotypic rescue of trichome branching. Each data set displays the percentages of the total trichome population with 2, 3, 4, 5 or 6 branches in plants grown in LD for 14 d. (E) Relaxation of 26S proteasome subunit hyperaccumulation in rpt2a-2 seedlings. Crude protein extracts were prepared from plants grown in LD for 10 d and subjected to immunoblot analysis with antibodies against RPT2, RPN1, RPN12a, and PBA1. Equal protein loading was confirmed by probing the blots with antibodies against Rubisco.

Figure 2.C-terminally tagged RPT2 subunits can rescue rpt2a mutant phenotypes. The RPT2a:RPT2a-T7, RPT2a:RPT2b-Myc, RPT2b:RPT2a-T7 and RPT2b:RPT2b-Myc transgenes were generated, plants transformed, and homozygous lines identified as in Figure 1. (A) Phenotypic rescue of inflorescence development. Wild-type (WT), rpt2a-2 and the four rescued rpt2a-2 lines expressing RPT2a-T7 or RPT2b-Myc under the control of either the RPT2a or RPT2b promoter were grown in LD for 7 weeks. (B) Phenotypic rescue of abnormal leaf shape. Pictured are newly expanded leaves taken from plants grown in LD for 7 weeks. (C) Phenotypic rescue of primary root growth. Each bar represents the average root length (± SD) of 20 plants grown in LD for 14 d. (D) Phenotypic rescue of trichome branching. Each data set displays the percentages of the total trichome population with 2, 3, 4, 5 or 6 branches in plants grown in LD for 14 d. (E) Relaxation of 26S proteasome subunit hyperaccumulation in rpt2a-2 seedlings. Crude protein extracts were prepared from plants grown in LD for 10 d and subjected to immunoblot analysis with antibodies against the T7 or Myc tags or the RPN1, RPN12a, and PBA1 proteins. Equal protein loading was confirmed by probing the blots with antibodies against Rubisco.

The N-terminal End is Important for Some Aspects of RPT2 Activity

Although the C-terminal HbYX motif of RPT2 has been intensively studied, less attention has been paid to the N-terminus, which possesses a potential N-myristoylation site that is conserved across many species, including humans, yeast and Drosophila.¹² Mutation of the critical glycine residue within the MGXXX(S/T) consensus sequence did not diminish the ability of an RPT2a transgene to rescue rpt2a-2 mutant phenotypes.¹² However, we show here that deletion of the N-terminal 13 amino acids of RPT2a attenuates this capacity. Expression of the RPT2:RPT2-ΔN13 transgene was again confirmed by the restoration of RPT2 protein levels to that seen in wild-type plants (Fig. 3E). Additionally, proteasome stress was alleviated, as shown by reduced accumulation of RPN1, RPN12a and PBA1 (Fig. 3E), and the stunted growth of the rpt2a-2 mutant was rescued (Fig. 3A). However, several other phenotypes, including lanceolate leaves, shortened root length, and increased trichome branching, remained compromised, indicating that the N-terminal region plays a role in these processes (Figs. 3B–3D). Given that mutation of the critical glycine residue within the N-myristoylation site rescued the characteristic rpt2a phenotypes, we conclude that the N-terminal region itself, rather than the specific glycine possibly involved in myristic acid attachment,²⁰ plays a crucial role in these developmental processes.

Figure 3. The N-terminus of Arabidopsis RPT2a is important for full functional activity. The RPT2a:RPT2a-ΔN13 transgene was generated, plants transformed, and homozygous lines identified as in Figure 1. (A) Phenotypic rescue of inflorescence development. Wild-type (WT), rpt2a-2, and three independent homozygous rpt2a-2 rescue lines expressing RPT2a-ΔN13 were grown in LD for 7 weeks. (B) Phenotypic rescue of abnormal leaf shape. Pictured are newly expanded leaves taken from plants grown in LD for 7 weeks. (C) Phenotypic rescue of primary root growth. Each bar represents the average root length (± SD) of 20 plants grown in LD for 14 d. (D) Phenotypic rescue of trichome branching. Each data set displays the percentages of the total trichome population with 2, 3, 4, 5 or 6 branches in plants grown in LD for 14 d. (E) Relaxation of 26S proteasome subunit hyperaccumulation in rpt2a-2 seedlings. Crude protein extracts were prepared from plants grown in LD for 10 d and subjected to immunoblot analysis with antibodies against RPT2, RPN1, RPN12a, and PBA1. Equal protein loading was confirmed by probing the blots with antibodies against Rubisco.

Phenotypes of Various 19S RP Mutants are Regulated in a Light-dependent Manner

In addition to the morphological phenotypes previously described, we discovered that mutations in RPT2a and several other RP subunit genes block the accumulation of core and linker histones (H1, H2B and H3),¹² which are required for robust nucleosome assembly.²¹ Treatment with the proteasome inhibitor MG132 restored histone levels to those seen in wild-type plants,¹² suggesting that disruption of the RP leads to increased histone degradation, either by the CP alone or in concert with alternative proteasome activators such as Blm10/PA200.^22,23 In agreement with a role in regulating histone dynamics, many phenotypes of rpt2a plants are shared by mutants in chromatin assembly factor 1 (CAF1), a heterotrimeric histone chaperone that delivers histones H3 and H4 onto DNA in the initial event of chromatin formation.²⁴ An attractive hypothesis is that the pleiotropic phenotypes of rpt2a result, at least partially, from reduced histone levels altering nucleosome dynamics, thus leading to altered DNA repair and overall gene expression, and compromised DNA replication and subsequent changes in ploidy.

While several RP subunit mutants, including rpn1a-4, rpt2a-3, rpn10–1 and rpn12a-1, have reduced histone accumulation, not all aberrant phenotypes are shared among these mutants. For instance, whereas trichome branching in rpn10–1 and rpn12a-1 mutants are similar to that of wild-type, it is significantly increased in rpn1a-4 and rpt2a-3 plants.¹² This suggests that the varied mutant phenotypes might not only arise from inadequate supplies of 26S proteasomes or altered histone levels, but from the unique functions of each subunit. To further investigate common features among RP mutants, we studied their growth under various light regimes, namely continuous light (CL; 24 h light), long-day (LD; 16 h light / 8 h dark) or short-day (SD; 8 h light / 16 h dark) photoperiods. Homozygous rpn1a-4, rpt2a-3, rpn10–1 and rpn12a-1 seedlings displayed strong growth abnormalities when exposed to CL and noticeably milder defects under LD (Fig. 4, upper and middle panels, respectively). However, when grown in SD, all four mutant lines closely resembled wild-type (Fig. 4, lower panel). Such photosensitivity had been previously reported for the rpn10–1 mutant,²⁵ which based on prior observations in yeast was speculated to reflect a challenged DNA repair pathway involving both RPN10 and RAD23, a UBL/UBA domain-containing factor that shuttles ubiquitylated proteins to RPN10 bound to the 26S proteasome.²⁶ Whether such a mechanism more generally involves other RP subunits is unclear.

Figure 4.Light-dependent changes in Arabidopsis seedling morphology in RP mutants with reduced 26S proteasome levels. Wild-type (WT), rpt2a-3, rpn1a-4, rpn10–1, and rpn12–1 mutant seeds were vapor-phase sterilized, vernalized in the dark at 4°C for 3 d, and germinated on MS medium containing 1% sucrose, 0.05% 2-(N-morpholino)ethanesulphonic acid (pH 5.7) and 0.75% agar. Seedlings were grown for 10 d under continuous light (CL), or long-day (LD; 16 h light / 8 h dark) or short-day (SD; 8 h light / 16 h dark) photoperiods. A representative seedling is shown in each case. Scale bar is 0.2 cm.

This observation suggests that the requirements for individual RP subunits, either within or independent of the 26S proteasome, might depend upon the light conditions to which the plant is exposed. Indeed, a recent study into the effects of high light intensity revealed that RP subunit transcript and protein levels are increased when plants are exposed to high fluence rates, in a manner at least partially dependent on ANAC078/NTL11, a member of the NAC transcription factor family.²⁷ Despite this, total proteasome activity is actually reduced when seedlings are exposed to high light fluences, possibly due to increased cellular levels of reactive oxygen species in response to the light stress.²⁷ While this study dealt with effects of light fluence rather than changes in photoperiodicity, it nevertheless indicates that the 26S proteasome is affected by light exposure.

One simple possibility is that plants require proteasome activity to respond to photodamage in LD or CL conditions, but that such damage is less prevalent in SD, hence impaired proteasome capacity is less disruptive to plant growth. However, it is also well established that regulated proteolysis plays a crucial role in many light-responsive signaling pathways,²⁸ with several factors being degraded by the 26S proteasome in response to exposure to or concealment from light. Examples include phytochromes A and B and the various phytochrome interacting factors (PIFs), which are stable in darkness but are rapidly degraded upon light irradiation.^28,29 While phytochromes and PIFs may not be the only factors stabilized in this case, they illustrate the potential role of the 26S proteasome in light-mediated signaling, and suggest a mechanism by which the severity of the proteasome mutants varies under different light conditions.