Coordination and assembly of protein complexes encoded across mitochondrial and nuclear genomes is assisted by CLPP2 in Arabidopsis thaliana

Protein homeostasis in eukaryotic organelles and their progenitor prokaryotes is regulated by a series of proteases including the caseinolytic protease (CLPP). CLPP has essential roles in chloroplast biogenesis and maintenance, but the significance of the plant mitochondrial CLPP remains unknown and factors that aid coordination of nuclear and mitochondrial encoded subunits for complex assembly in mitochondria await discovery. We generated knock-out lines of the single gene for the mitochondrial CLP protease subunit, CLPP2, in Arabidopsis thaliana. Mutants had higher abundance of transcripts from mitochondrial genes encoding OXPHOS protein complexes, while transcripts for nuclear genes encoding other subunits of the same complexes showed no change in abundance. In contrast, the protein abundance of specific nuclear-encoded subunits in OXPHOS complexes I and V increased in CLPP2 knockouts, without accumulation of mitochondrial-encoded counterparts in the same complex. Protein complexes mainly or entirely encoded in the nucleus were unaffected. Analysis of protein import, assembly and function of Complex I revealed that while function was retained, protein homeostasis was disrupted through decreased assembly, leading to accumulation of soluble subcomplexes of nuclear-encoded subunits. Therefore, CLPP2 contributes to the mitochondrial protein degradation network through supporting coordination and assembly of protein complexes encoded across mitochondrial and nuclear genomes. One sentence summary CLPP contributes to the mitochondrial protein degradation network through supporting coordination and assembly of protein complexes encoded across mitochondrial and nuclear genomes.


Introduction
photosynthesis (Dong et al., 2013). Therefore, the plant CLPPR protease complex 141 appears to play an essential role in plastid and chloroplast biogenesis and development.

143
Arabidopsis mitochondria contain a homotetradecameric CLPP2 core protease encoded 144 by a single nuclear gene and three CLPX chaperones (CLPX1-3) (Peltier et al., 2004). The 145 CLPP2 core has been detected by Blue native PAGE with a molecular mass of ~320 kDa 146 (Senkler et al., 2017), which is consistent with a predicted tetradecamer structure (Peltier 147 et al., 2004), but no T-DNA lines for CLPP2 exist. In this study, we developed two knock- The serine-type, ATP-dependent CLP protease system in Arabidopsis mitochondria 159 contains a CLPP2 core subunit (AT5G23140) and three homologous CLPX chaperone 160 subunits (CLPX1, AT5Gg53350; CLPX2, AT5G49840, CLPX3, AT1G33360) (van Wijk, 161 2015). So far, there is no reported or available T-DNA insertion line within the coding 162 region for CLPP2 (AT5G23140). In order to acquire independent stable CLPP2 mutants, 163 we used a CRISPR-Cas9 guided system to knock-out the CLPP2 core subunit 164 (Supplemental Figure 1), resulting in two individual CRISPR-Cas9 mutants, clpp2-1 and 165 clpp2-2, showing a complete knock-out of the target gene. Both mutations occurred within 166 exon 1 as shown by guides, primers and restriction sites/enzymes ( Figure 1A).

168
The first mutant clpp2-1 showed two separate genetic events that caused the CLPP2 169 knock-out: firstly, a single adenine insertion at position of 25bp, introduced a frame shift 170 and disrupted the restriction site of NlaIV at 24bp ( Figure 1B); secondly, the introduction 171 of the adenine at CRISPR site 1 created an identical 3' end as the 5'end at CRISPR site 3. 172 It is likely that this caused a second error from the nonhomologous end-joining (NHEJ) 173 DNA repair mechanism or from the alternative nonhomologous end-joining DNA repair 174 mechanism (MMEJ) known for insertions, deletions and inversions (McVey and Lee, 2008), 175 inserting the complete fragment between CRISPR site 1 and CRISPR site 3 in a reverse 176 complemented orientation. Consequentially, five downstream stop codons (TGA or TAA) 177 were generated and the second NlaIV restriction site was shifted to the position at 95bp 178 ( Figure 1B). The second mutant clpp2-2 had a single NHEJ error at CRISPR site 2 with 179 an adenine insertion at the position of 124bp, resulting in a disruption of the NruI restriction 180 site and the introduction of 3 downstream stop codons (TAA or TGA) ( Figure 1B).

182
We confirmed the disruption of CLPP2 in the mutants in the genetic sequence as well as 183 its consequence on mRNA and protein levels. As the two mutations resulted in disruption 184 of restriction sites within exon 1, mutants and WT genomic DNA was amplified using 185 primers F1 and R1 ( Figure 1A) and digested with NlaIV for clpp2-1 and NruI for clpp2-2, 186 respectively. NlaIV digested WT PCR product resulted in three fragments, a long 165bp Based on RNAseq analysis, the transcript levels of CLPP2 were significantly reduced in 198 clpp2-1 and clpp2-2 to about 15% of the WT level ( Figure 1D). The expression level of 199 representative house-keeping genes such as general TUBULIN6 (AT5G12250) and 200 mitochondrial ATVDAC1 (AT3G01280) were unaffected in both mutants ( Figure 1D). 201 While CRISPR-Cas9 systems shouldn't actively alter transcript levels, the introduced 202 frameshifts forming several stop codons into the sequence of CLPP2, more than 50bp 203 away from the exon-exon junction (Figure 1B), created a high confidence position for 204 exon-exon junction nonsense-mediated mRNA decay which may contribute to the 205 observed low transcript abundance (Lloyd, 2018). 206 To detect the protein abundance of CLPP2, we isolated mitochondria and conducted 207 multiple reaction monitoring (MRM)-based targeted proteomic analysis and included the 208 theoretical peptide of the inverted CLPP2-1 section. The peptide MRM transitions were 209 presented in Supplemental Table 1. CLPP2 protein abundance in mitochondria from 210 both mutants was undetectable, while all selected peptides of CLPP2 were detected in WT 211 ( Figure 1E). We also did not detect any evidence of theoretical CLPP2-1 peptides. Two To evaluate global transcript abundance changes upon the loss of the mitochondrial 222 CLPP2, we used high-throughput RNA-sequencing to find and quantify differentially 223 expressed genes (DEGs). We analysed the transcriptome of hydroponically grown 224 seedlings of clpp2-1, clpp2-2 and WT, with three biological replicates for each genotype.

225
Based on a filter of a log2 fold change (log2FC) exceeding ±0.4 and adjusted p-values ≤ 226 0.05, we identify 62 DEGs using the R package Deseq2 (Love et al., 2014), which showed 227 a consistent pattern in both mutants (Figure 2). Only 0.1% of nuclear gene transcripts had 228 significant changes in abundance in both clpp2-1 and clpp2-2. This was considerably 229 lower than the 33% of mitochondrial transcripts that showed significant changes in 230 abundance in both mutants (Figure 2), indicating a specific impact of the disruption of 231 CLPP2 on the abundance of transcripts from mitochondrial genes.

232
In clpp2-1 and clpp2-2, 41 of the 52 upregulated DEGs from nuclear and mitochondrial 233 genomes encode mitochondrial-localised proteins, while 3 of 8 downregulated DEGs, 234 other than CLPP2, encodes a mitochondrial-localised protein ( Table 1). Two genes that 235 encode non-mitochondrial targeted proteins and have unknown function stand out, 236 showing very high and consistent log2FC values. AT4G36850 encodes a conserved PQ-

252
Loss of CLPP2 alters mitochondrial protein homeostasis 253 Changes in transcript abundance often show only a poor correlation with the abundance of 254 finally accumulated proteins (Haider and Pal, 2013). To identify differentially expressed 255 proteins (DEPs) in isolated mitochondria from hydroponically grown Arabidopsis seedlings, 256 we conducted a quantitative proteomic approach. We detected 797 proteins from 257 mitochondrial membrane fractions in mutants and WT (See Supplemental Table 2). 258 Among them, we identified 22 DEPs with increased abundance and 5 DEPs with 259 decreased abundance. These showed a consistent pattern in clpp2-1 and clpp2-2 with a 260 log2 fold change (log2FC) exceeding ±0.4 and p-values ≤ 0.05 (Figure 3). Details of the 261 27 DEPs are listed in Table 2 and are grouped based on functional categories. We 262 detected the accumulation of ATP2 (AT5G08690) for Complex V and four subunits of 263 Complex I (24kDa: AT4G02580, 51kDa: AT5G08530, 75kDa: AT5G37510 and B14: 264 AT3G12260) ( Table 2). Three of the over-accumulated Complex I subunits belong to the 265 N-module of the matrix arm, which is assembled before it is attached to the membrane 266 arm of Complex I (Ligas et al., 2019). We observed an accumulation of proteins related to 267 protein synthesis, such as three components of the mito-ribosome (AT4G3090, 268 AT1G61870 and AT5G64670) and the putative RNA helicase AT3G22130 (RH9) as well 269 as mitochondrial proteases, such as MPPα-1,α-2; MPPβ, CLPX-1 and PREP1 (Table 2). 270 MPPs are mitochondrial peptidases that are embedded in Complex III and are responsible 271 for mitochondrial presequence cleavage after protein import (Braun et al., 1992;Zhang et 272 al., 2001). CLPX-1 is the chaperone subunit of the CLPXP complex (van Wijk, 2015) and 273 PREP1 degrades mitochondrial pre-sequences after their cleavage from proteins (Kmiec 274 and Glaser, 2012;Kmiec et al., 2014). A change to the abundance of these proteases in  Table 2).

282
To investigate a potential modification in pre-sequence processing or partial degradation of 283 mature proteins, we conducted a combined approach of protein labelling with isobaric tags 284 for relative and absolute quantification (Ross et al., 2004) and charge based fractional 285 diagonal chromatography (ChaFRADIC) (Venne et al., 2013;Venne et al., 2015) to 286 analyse the N-terminal mature sequence of mitochondrial proteins. Sequence logo 287 analysis indicated that there was no difference in pre-sequences cleavage as there were 288 very conserved -2R and -3R cutting sites detected in mitochondrial proteins in both 289 mutants and WT (Supplemental Figure 3A & 3B). There was also no apparent difference 290 in the abundance of any pre-sequence peptides between mutants and WT (Supplemental 291 Figure 3C). In addition, no significant difference was observed in cleavage sites in the 292 middle or C-terminus of protein sequence between the two mutants and WT 293 (Supplemental Figure 3C & 3D). Therefore, loss of CLPP2 had no measurable impact on 294 mitochondrial protein maturation and left no partially degraded parts of mitochondrial 295 proteins that we could detect. To directly compare between mitochondrial transcript and protein abundance changes for 301 individual subunits located in the same protein complex, we compiled the changes in 302 transcript level and protein abundance for all subunits of the mitochondrial OXPHOS 303 system and the mitochondrial ribosome (Figure 4).
Complex I, the largest protein 304 complex of the mitochondrial OXPHOS system, is encoded by nine mitochondrial encoded 305 genes and 47 nuclear encoded genes (Ligas et al., 2019;Meyer et al., 2019). Our data 306 showed a consistent transcript upregulation for five mitochondrial encoded subunits (NAD1, 307 3, 5, 7) in both mutants, but we did not find NAD7 protein accumulation in either mutant 308 ( Figure 4A). In contrast, complex I subunits (24 kDa, 51 kDa, 75 kDa and B14), encoded 309 by four nuclear genes, were highly accumulated in protein abundance with no change in 310 transcript levels ( Figure 4A). A similar pattern was observed for Complex V; five (ATP1, 311 ATP6 (AT3G46430, AT5G59613), ATP7, ATP8) out of six mitochondrial transcripts had 312 higher abundance in both mutants, but ATP1 and ATP8 did not show any accumulation at 313 the protein level ( Figure 4E). There was no consistent transcript upregulation of nuclear 314 genes encoding Complex V subunits in both mutants, but there was high accumulation of 315 the ATP2 protein, encoded by the nuclear genome, in both mutants ( Figure 4E).

317
Complex III contains only one mitochondrial encoded subunit (Braun and Schmitz, 1992;318 Meyer et al., 2019) and its expression appeared to be unaffected transcriptionally in both 319 mutants. Changes in protein abundance of MPPα-1, MPPα-2 and MPPβ, were observed 320 but without any changes in transcript level for these nuclear encoded genes ( Figure 4C).

321
Complex IV has three mitochondrial subunits, all showed upregulation in transcript 322 abundance in both mutants compared to the WT, and 13 nuclear genes encoded subunits 323 of Complex IV showed no change in transcript or protein abundance ( Figure 4D).

324
Complex II of the OXPHOS system contains only nuclear-encoded subunits and appeared 325 to not be affected by a disruption of the mitochondrial CLPP2 protease at the transcript or 326 protein level ( Figure 4B). We did not find consistent transcriptional response in both 327 CLPP2 mutants for mitochondria-encoded subunits of the mito-Ribosome ( Figure 4F), yet 328 two nuclear-encoded ribosomal large subunits (AT4G30930, AT5G64670) and one 329 nuclear-encoded ribosomal PPR336 (AT1G61870) were increased in protein abundance in 330 both mutants ( Figure 4F).

332
Loss of CLPP2 impacts protein homeostasis and the assembly of Complex I

333
Changes in the transcript level and protein abundance of specific Complex I subunits lead 334 us to assess the impact on respiratory chain Complex I in both mutants as an exemplar of 335 a complex affected in CLPP2 mutants. We firstly analysed Complex I activity in both 336 mutants compared to WT as the rate of de-amino NADH-dependent FeCN reduction but 337 did not find any difference in enzymatic activity ( Figure 5A). To determine the abundance 338 of assembled Complex I, we separated mitochondrial complexes using BN-PAGE and 339 stained gels to visualise proteins with Coomassie ( Figure 5B). We did not observe any subunit in clpp2-1 and clpp2-2 was observed when compared to WT ( Figure 5D).

357
However, in the soluble fraction the accumulation was 4-6.5-fold and significant for all 3 358 subunits ( Figure 5D). This indicated that the accumulation of specific Complex I N-module 359 subunits appeared to be an accumulation of an unassembled subcomplex even though the 360 steady-state intact Complex I abundance was not affected in either mutant.

362
We conducted an in-vitro import assay of a radio-labelled 24 kDa subunit to understand 363 the nature of the accumulation of the upregulated Complex I subunits in clpp2. There was 364 no apparent difference in the degree of import of the 24 kDa subunit in WT and clpp2-1 on 365 SDS-PAGE ( Figure 5E). Next, we tested for any impairments of Complex I assembly in 366 clpp2-1 by analysis of radiolabelled imports of the 51 kDa and 23 kDa subunit after 2 and 367 16 hours using blue native-PAGE ( Figure 5F). After 2 hours of 51 kDa import and 368 assembly, we found the emergence of new fully assembled Complex I and supercomplex 369 I-III 2 in WT and clpp2-1, with a higher abundance in WT ( Figure 5F). After 16 hours of 51 370 kDa import and assembly, we observed the same fully assembled Complex I and 371 supercomplex I-III 2 bands, however they were a lot less abundant ( Figure 5F), which could 372 be due to degradation of the imported and assembled protein over time. After two hours of 373 a 23 kDa subunit import we observed similar radiolabelled Complex I and supercomplex I-

374
III 2 bands in WT and clpp2-1, again displaying higher abundances in the wild type ( Figure   375 5F). After 16 hours, we saw the fully assembled Complex I and Supercomplex I-III 2 bands 376 in wild type and the mutant, but they were relatively similar in abundance to the two hour 377 import and assembly assay, which could be an indication of less degradation than for the

416
While it has been long known that a CLPP exists in plant mitochondria (Peltier et al., 2004), 417 there was no information on its function due to lack of genetic resources for its disruption 418 and analysis. In this study, we successfully developed two mutant lines using CRISPR-

419
Cas9 to knockout the Arabidopsis mitochondrial CLPP2 gene to provide direct insight into 420 its function not only in mitochondrial protein homeostasis but also in coordinated regulation  Table   499 3). subunits phenomenon in a transition from starved to re-fed Arabidopsis callus culture cells, 503 we found here that carbon starvation did not produce a phenotype in clpp2 plants 504 (Supplemental Figure 4), but transient response to nitrogen starvation did (Fig 6). This 505 implies that non-stoichiometric accumulation of unused subunits is not in itself egregious 506 for plant mitochondrial function, but perhaps it is suboptimal as is shown by the decreased 507 Complex I assembly when N-module accumulates in clpp2 (Figure 5, Figure 7). The 508 reasons for the specific need for mixed subunit complexes to turnover in this way are 509 unknown. However, there is additional evidence for such maintenance of Complex I. The containers were seated on a rotating shaker in the long-day conditions as described above 568 and seedlings were harvested two weeks after germination. were obtained per sample, which were mapped to Arabidopsis reference genome Ensembl, 580 v34 and quantified with Salmon (v0.14.0) (Patro et al., 2017). Differential analysis was 581 performed with the R package DESEQ2 (Love et al., 2014).and transcripts with a log2 fold 582 change between mutant and WT (log2FC) exceeding ±0.4 and an adjusted p-value less 583 than or equal to 0.05 were considered differentially expressed.   Three hundred µg mitochondrial proteins were dissolved with 5mg digitonin/mg protein and 689 then separated with 4.5-16% gradient BN-PAGE gel (Schertl and Braun, 2015 identified using an Agilent 6550 Q-TOF as described in (Nelson et al., 2014).  Images were acquired using a Gatan Orius CCD camera.     Cole Clpp2-1: WT shows the exon 1 sequence map with the location of two NlaIV restriction sites. Clpp2-1 has an adenine insertion (A) at CRISPR site 1 disrupting (X) one NlaIV restriction site and an insertion of the reverse complemented wild type sequence between CRISPR site 1 and CRISPR site 3 introducing 5 stop codons (TAA, TGA) resulting in change of the second NlaIV restriction site. Clpp2-2: WT shows the exon 1 sequence map with the location of one NruI restriction site. Clpp2-2 has an adenine insertion (CRISPR site 2) resulting in the disruption of the NruI restriction site and insertion of 3 stop codons. (C) Confirmation of CRISPR-Cas9 gene disruption based on PCR fragment restriction digest amplified by F1 and R1 primers. NlaIV and Nrul digestion are shown in left and right, respectively. (D) Confirmation of CRISPR-Cas9 gene disruption at the transcript level based on RNA seq analysis. The normalised counts of the target gene CLPP2, and the two control genes Tubulin6 and ATVDAC1 are presented. ** represent p ≤ 0.01 (n = 3). (E) Confirmation of CRISPR-Cas9 gene disruption at the protein level based on a targeted proteomic approach. Fold changes of the target protein CLPP2, and the two control proteins mtHSP70 and ATVDAC2 are presented. *** represent p ≤ 0.0001 (n = CLPP2 : 3 peptides, 2 replicates, mtHSP70 : 3 peptides, 2 replicates, ATVDAC2 : 2 peptides, 2 replicates). The volcano plot illustrates overlapping significant changes of total transcript abundances of the two KO-mutants compared to the wild type. The negative Log10 transformed adjusted p-values (p.adj) are plotted against the log2 fold change (log2FC) of the transcript abundances measured by RNAseq. Dashed lines separate gene transcripts with significant changes in abundance with the log2FC cut off of ±0.4 and adjusted p-value cut off of 0.05. Depleted gene transcripts are shown in red and accumulated gene transcripts in green. Transcripts with p.adj values under 1x10 -15 are displayed as arrow shapes. Transposable elements and hypothetical proteins are displayed as empty circles. The inserted table displays the summary total number of detected nuclear and mitochondrial genes and the overlapped numbers with significant changes in abundance from both mutants against the wild type.   Figure 2A) and protein abundances (protein listed at the bottom) from quantitative proteomics ( Figure 3A). The subunits with accession number and abbreviations from individual complex are listed on the left side. In heatmaps, the colour code represents the log2 fold change of the individual mutants compared to the wild type with green for accumulation and red for depletion in the mutants. Grey colour indicates un-detected proteins. * represent p ≤ 0.05, ** represent p ≤ 0.01, *** represent p ≤ 0.001). (n= 3 for transcript and n = 4 for protein).

Figure 5: Changes in protein abundance for Complex I N-module subunits and
Complex I enzymatic activity and in-vitro import for Complex I assembly in CLPP2 mutants. (A) Complex I enzyme activities in isolated mitochondria from mutants and wild type using de-amino NADH as substrate and FeCN as donor measured using photospectrometer assay (n=5). (B) BN-PAGE of isolated mitochondria of WT, clpp2-1 and clpp2-2, stained with Coomassie blue was shown in the left and the image of BNnative in gel complex I activity stain (NADH and NBT) was presented in the right. Major OXPHOS components are marked on the gel for size comparison. Bands of interest are marked with a (Supercomplex I-III2), b (Complex I) and c (potential Complex I assembly intermediate). (C) Mass spectrometry evidence for N-module proteins (24KDa and 51 KDa) present in band c in the BN-PAGE. (D) Protein abundances of three Complex I matrix arm N-module subunits (24 Kda, 51 kDa and 75kDa) in membrane and soluble fractions from isolated mitochondria. Data is shown as fold change in peptide abundance compared to the wild type median. ** represent p ≤ 0.01 (n= 4). (E) In-vitro import of radio-labelled 24kDa subunit of Complex I into isolated mitochondria of WT and clpp2-1 at 5, 10-and 15-minutes incubation time. (F) In-vitro import of radio-labelled Complex I 51KDa subunit (left) and 23KDa subunit into isolated mitochondria of WT and clpp2-1 at 2 hours and 16 hours incubation time. The blue-native (BN) gels were imaged for radioactive intensity. Major OXPHOS components are marked on the gel for size comparison.  In both mutants, the subunits in unassembled N-module subcomplex accumulated at much higher level than these in the assemble holo-complex as indicated by targeted proteomic analysis using MRM in soluble and membrane fractions. With interference of pre-existence of highly (~5-fold) accumulated N-module subunits in matrix, the assembly of imported 51 KDa subunit and 23 KDa subunit (with physical interaction with N-module subunits) into assembled Complex I and supercomplexI+III were reduced ( Figure 5F).