Comparative proteomic study of Arabidopsis mutants mpk4 and mpk6

Abstract

Arabidopsis MPK4 and MPK6 are implicated in different signalling pathways responding to diverse external stimuli. This was recently correlated with transcriptomic profiles of Arabidopsis mpk4 and mpk6 mutants, and thus it should be reflected also on the level of constitutive proteomes. Therefore, we performed a shot gun comparative proteomic analysis of Arabidopsis mpk4 and mpk6 mutant roots. We have used bioinformatic tools and propose several new proteins as putative MPK4 and MPK6 phosphorylation targets. Among these proteins in the mpk6 mutant were important modulators of development such as CDC48A and phospholipase D alpha 1. In the case of the mpk4 mutant transcriptional reprogramming might be mediated by phosphorylation and change in the abundance of mRNA decapping complex VCS. Further comparison of mpk4 and mpk6 root differential proteomes showed differences in the composition and regulation of defense related proteins. The mpk4 mutant showed altered abundances of antioxidant proteins. The examination of catalase activity in response to oxidative stress revealed that this enzyme might be preferentially regulated by MPK4. Finally, we proposed developmentally important proteins as either directly or indirectly regulated by MPK4 and MPK6. These proteins contribute to known phenotypic defects in the mpk4 and mpk6 mutants.

MAPKs are important signalling molecules which are involved in transduction of signal derived from many environmental and developmental stimuli¹. MAPK signalling results in gene activation triggering appropriate defense responses or in the activation or repression of developmentally-regulated proteins modulating plant growth and development². The complete sequencing and annotation of the Arabidopsis genome resulted in the identification of 20 MAPK-encoding genes, while MPK4, MPK6 and MPK3 are the most studied ones and they play dominant roles in the transduction of stress and developmental signal as well as in the cross-talk with other signaling pathways².

MPK4 and MPK6 are commonly activated by cold and salt stress downstream of MEKK1 (MAP3K) and MKK2 (MAP2K) in Arabidopsis³, while MPK4 via this pathway also negatively regulates salicylic acid (SA) and ROS production⁴. MEKK1-mediated activation of MPK6 (together with MPK3) was reported also via MKK4/MKK5 in the response to bacterial elicitor flagellin. This signaling cascade affects expression of defense related genes⁵. The MKK4-MPK6 module also phosphorylates and activates ACC synthases ACS2 and ACS6 which are key enzymes for ethylene synthesis⁶. Additionally, MPK6 is specifically activated by MKK3 in response to jasmonic acid (JA) and negatively regulates the JA signalling pathway⁷. MPK6 is also activated by the pathway that includes ANP1 (MAP3K)–MKK4/MKK5 upon hydrogen peroxide treatment⁸. In contrary, MPK4 is activated by MEKK1 (MAP3K)–MKK2 (MAP2K) pathway under oxidative stress⁹.

Along with roles of MPK4 and MPK6 in various stress responses, they are involved in multiple developmental processes. Thus, MPK6 is involved in embryogenesis¹⁰, stomata formation¹¹, cell division plane orientation and root development¹² and it is localized in the secretory Trans-Golgi network (TGN) vesicles and plasma membrane. MPK4 controls cell division, root growth and formation of root hairs¹³,¹⁴.

An absence of particular members of the MAPK cascade in Arabidopsis single and double knockout mutants can lead to considerable molecular, physiological, developmental and phenotypic changes, and it can modify their stress responses³,¹⁰. However, there is only very scarce information about the proteome wide effects resulting from genetically-affected MAPK signaling in plants¹⁵.

Roots of mpk4 mutant have altered organization of cortical microtubules leading to radial root expansion and adverse effects on root hair morphogenesis¹³. At the subcellular level, the mpk4 mutant showed aberrant spindle and phragmoplast formation and drastically delayed or abortive mitosis and cytokinesis¹⁴. Additionally, hydrogen peroxide accumulation, reduced expression of auxin inducible marker genes and elevated levels of mRNA encoding stress response proteins was observed in the seedlings of mpk4 mutant⁴. Detailed analyses of mpk6 knock-out mutants showed root phenotypes, which are consequences of ectopic cell divisions and aberrant orientation of cell division, resulting in disordered root cell files in the mpk6 mutants¹². Here, we have studied major changes associated with MPK4 and MPK6 deficiency in Arabidopsis roots on the proteome level. We aimed to find protein candidates contributing to specific root phenotypic features of mpk4 and mpk6 mutants. In addition, we also focused on proteome changes associated with oxidative stress and defense responses in these two mutants.

Results

Overview of root proteomes of Arabidopsis mpk4 and mpk6 mutants

We performed proteomic analyses of mpk4 and mpk6 mutant roots in direct comparison to the Col-0 wild type. Representative pictures of analyzed mutants are provided in Fig. 1. This shot gun MS/MS-based proteomic analysis identified in average 437 proteins in the mpk4 and 444 proteins in the mpk6 mutant. All proteins with statistically significant changes in their abundances between wild type and mutant plants were selected using one-way Anova test. They are provided in Supplementary Tables S1 and S2.

Figure 1.

Representative pictures of 14 days old wild type plants Col-0 (A), mpk4 (B) and mpk6-2 short (C) and long root (D) mutants used for proteomic and biochemical analyses. Bar: 1 cm.

First, we compared the numbers of differentially regulated proteins in both mutants (Supplementary Table S3). Only proteins with change higher than 1.5 fold were considered in this study. We detected 63 differentially abundant proteins in the mpk4 mutant when compared to Col-0 wild type seedlings, while there were 32 differentially abundant proteins in the mpk6 mutant. In addition, several proteins (33 and 19 for mpk4 and mpk6, respectively) were detected only in one sample, either in the wild type, or in the one of the mutants, indicating that in such cases their abundance was below the detection sensitivity limit. Altogether, we have found 96 and 51 differentially abundant proteins in the roots of mpk4 and mpk6 mutants, respectively. These data might indicate that MPK4 deficiency has broader influence on Arabidopsis basal root proteome as compared to MPK6.

In order to functionally classify these proteins in differential proteomes of both mutants, we performed analyses of known and predicted functional protein association networks using STRING¹⁶ (Supplementary Figs S1 and S2) and gene ontology (GO) annotation according to biological process, molecular function and cellular compartment using Blast2Go¹⁷ software (Fig. 2A–C). These analyses revealed that mpk4 and mpk6 mutants did not substantially differ in the number of GO annotations detected in differential proteomes according to biological function at the 3^rd level of ontology (Fig. 2A). Nevertheless, some interesting differences, including different number of proteins in the two mutants, have been found in GO annotations in terms of response to stress, cellular component organization and single multicellular organism process (Fig. 2A) as well as in terms of molecular functions (Fig. 2B). On the other hand, no changes were detected in GO annotations in terms of subcellular localization (Fig. 2C).

Figure 2.

Comparison of gene ontology classification (at 3^rd level) of mpk4 and mpk6 mutant root differential proteomes according to (A) biological process, (B) molecular function, (C) cellular compartment. Arrowheads indicate most prominent differences between the mutants.

We also performed a GO annotation analysis on whole proteomes of mpk4 and mpk6 roots (Supplementary Figs S3–S5). In this case, only minor differences were observed between whole proteomes of the mutants.

Next, we searched for proteins which abundances were commonly changed in both mutants compared to wild type (Supplementary Table S4). Eight differentially abundant proteins were detected in both mpk4 and mpk6 roots. Two of them showed decreased abundance in both mutants (arabinogalactan protein 31 and profilin 2) and three other proteins had increased abundance in both mutants (glyceraldehyde 3-phosphate dehydrogenase, TRAF-like family protein and translational initiation factor 4A-1). These small numbers of commonly regulated proteins show that root proteomes of the mutants are quite different.

Does MPK3 compensate for missing MPK6 in roots of the mpk6 mutant?

MPK3 and MPK6 show high level of functional redundancy in response to external stimuli⁸ and in regulating plant development¹⁸. It was also found that MPK3 activities might be increased in mpk6 mutants in response to various external stimuli. In this study, we evaluated the MPK3 abundance and activity in the roots of the mpk6 untreated mutant using affinity purified antibodies against MPK3 and phosphorylated mammalian ERK1/2 (phospho-p44/42, pERK) recognizing pTEpY-dual phosphorylation motif (Fig. 3), preferentially in MPK6 and MPK3. We detected increased abundance of MPK3 in the mpk6 mutant as compared to the Col-0 wild type (Fig. 3A,B). By using pERK antibody we detected two bands in Col-0 wild type roots, corresponding to phosphorylated MPK6 and MPK3, respectively. As expected, mpk6 mutant showed only a single band, corresponding to the activated MPK3 which was, however, less prominent in the mutant as compared to the Col-0 wild type (Fig. 3A,C). These data suggest that differences in the root proteome composition of the mpk6 mutant might be, at least partially, attributed also to the higher abundance but lower activity of MPK3 in the roots of this mutant.

Figure 3. Imunoblotting analysis of MPK3 activity in the mpk6 mutant using phospho-specific pERK antibody.

(A) Representative immunoblots of wild type and mpk6 mutant roots probed with anti-MPK3 (lane MPK3), anti-MPK6 (lane MPK6) and anti-phospho-p44/42 (pERK) antibodies. (B) Quantification of the band optical density corresponding to MPK3 in (A). (C) Quantification of the band optical density corresponding to phosphorylated MPK3 (pMPK3) (band of 43 kDa in lane pERK in A) in wild type and mpk6 mutant roots. Error bars are standard deviations calculated from 3 biological replicates. Stars indicate significant difference at p ≤ 0.05 according to Student t-test.

Prediction of putative MPK4 and MPK6 targets

Proteomic analyses performed on the loss-of-function and knockout mutants can help to identify both direct and indirect targets of MAPKs. The interaction of MAPKs with their direct protein targets is conditioned by the presence of MAPK docking site, MAPK-specific phosphorylation site and common subcellular localization with MAPKs. Therefore, we analyzed proteomically-identified proteins of mpk4 and mpk6 mutant roots for such motifs and gained information about localization of these putative target proteins. Proteins containing both docking and phosphorylation motifs are listed in Tables 1 and 2. Details of prediction including probabilities, scores, putative phosphorylation sites and amino acid sequences of the docking domains are provided in Supplementary Tables S5–S8. We have found 17 and 8 putative target proteins in the mpk4 and mpk6 mutants, respectively. The localization of some putative target proteins corresponded with the published localization of MPK4 and MPK6 in Arabidopsis¹²,¹⁴,¹⁹. These were for example elongation factor EF-2, uncharacterized protein (gi:240254562) and mRNA decapping complex VCS in the mpk4 mutant as well as CDC48A and phospholipase D alpha 1 in the mpk6 mutant, while heat shock protein 70–3 was found in both mutants.

Table 1. List of differentially regulated proteins containing both MAPK-specific phosphorylation site and MAPK docking site as predicted by GPS 3.0 software and Eukaryotic Linear Motif Resource (http://elm.eu.org/index.html) in the roots of the mpk4 mutant.

Accesion number	Protein name	Phosphopeptide sequences	Docking sequence	Fold change	Cell compartment
gi|145324054	arabinogalactan protein 31	EVNHKTQTPSLAPAP	KFNRSLVAV, 180–188 [A]	unique in wild type	cell wall72
		HPHPPAKSPVKPPVK	KKLGKSTVVV, 285–294 [A]
		PPVKAPVSPPAKPPV	KLGKSTVVV, 286–294 [A]
		PPVKPPVSPPAKPPV
		APVKPPVSPPTKPPV
		PPTKPPVTPPVYPPK
gi|15226573	Ferredoxin–nitrite reductase	SFSLTFTSPLLPSSS	KPKRSVLV, 19–26 [A]	unique in wild type	plastid (prediction)
			KIEREPMKL, 74–82 [A]
			KSSKDDIDVRL, 103–113 [A]
			RKWNVCV, 246–252 [A]
			KDGRFGFNLLV, 274–284 [A]
			KRCEEAIPL, 291–299 [A]
			RQKTRMMWL, 328–336 [A]
			KKGVRVTELVPL, 554–565 [A]
			KGVRVTELVPL, 555–565 [A]
gi|15233349	aconitate hydratase 1	RATIANMSPEYGATM	RIDKLPYSIRI, 35–45 [A]	unique in wild type	cytosol, mitochondria73
		YFKGMTMSPPGPHGV	KRPHDRVPL, 378–386 [A]
			KKACDLGL, 458–465 [A]
gi|18400212	dihydrolipoamide acetyltransferase, long form protein	TTSTKLSSPMAGPKL	RRDHAVAV, 20–27 [A]	unique in wild type	mitochondria74
		EIGMPSLSPTMTEGN
gi|22331076	Subtilase family protein	PLLLCFFSPSSSSSD	RRHPSVISV, 93–101 [A]	unique in wild type	extracellular (prediction)
		LLRSLPSSPQPATLL
		HGFSARLSPIQTAAL
		REIHTTHTPAFLGFS
		LGTLIGPSPPSPRVA
		LIGPSPPSPRVAAFS
		ANVEIDVSPSKLAFS
gi|240254562	uncharacterized protein	SSSGNVTTPTQTAST	KSRDIDLSF, 1264–1272 [A]	unique in wild type	nucleus (prediction)
gi|30682607	mRNA decapping complex VCS	PGISAQPSPVTQQQQ	RKAQPLVVL, 352–360 [A]	unique in wild type	cytoplasmic foci75
		TPPLNLQSPRSNHNP	KESKRLEVAL, 918–927 [A]
		TLPQLPLSPRLSSKL	KRLEVAL, 921–927 [A]
gi|15232845	probable mitochondrial-processing peptidase subunit beta	DSVPASASPTALSPP	RRSQRRLFL, 11–19 [A]	13.11	mitochondria
		SASPTALSPPPPHLM	RINRERDVIL, 210–219 [A]
gi|15221019	GDSL esterase/lipase	ITVAGQNSPVVALFT			extracellular (prediction)
		KFSDGLITPDFLAKF	KFMKIPLAI, 88–96 [A]	5.48
		REFWVPPTPATVHAS
gi|15232603	60S acidic ribosomal protein P0-2	KGTVEIITPVELIKQ	RKGLRGDSVVL, 44–54 [A]		ribosome
		YDNGSVFSPEVLDLT	KGLRGDSVVL, 45–54 [A]	4.61
			KINKGTVEI, 148–156 [A]
gi|186513287	argininosuccinate synthase	ALNGKALSPATLLAE	RGKLKKVVL, 93–101 [A]		plastid76
			KKHNVPVPV, 255–263 [A]	3.63
			KKDMYMMSV, 293–301 [A]
			KLYKGSVSV, 393–401 [A]
gi|15240765	voltage dependent anion channel 2	DITATLGSPVISFGA	KHPRFGLSLAL, 264–274 [A]	2.41	mitochondria, plasma membrane24
gi|15218090	putative mitochondrial-processing peptidase subunit alpha-1	VAETSSSTPAYLSWL			mitochondria
		LKIASETTPNPAASI	RKMKVEI, 196–202 [A]	2.31
		GYSGPLASPLYAPES
gi|334185190	heat shock protein 70-3	DLLLLDVTPLSLGLE	RARFEELNI, 305–313 [A]	0.57	nucleus, cytoplasm30
			RIPKVQQLLV, 348–357 [A]
gi|14532542	proteasome subunit alpha type-5-A	AVEKRITSPLLEPSS	KTKEGVVLAV, 41–50 [A]	0.32	peroxisome, cytoplasm (prediction)
gi|30691626	heat shock protein 70-1	GTSGTEQTPEAEFEE	KRSDNIDL, 289–296 [A]	0.32	cytosol77
			KKQLIDL, 573–579 [A]
gi|30696056	elongation factor EF-2	EEMQRPGTPLYNIKA	KRLAKSDPMVV, 509–519 [A]	0.08	cytoplasm78
		KGLKEAMTPLSEFED	RLAKSDPMVV, 510–519 [A]

Table 2. List of differentially regulated proteins containing both MAPK-specific phosphorylation site and MAPK docking site as predicted by GPS 3.0 software and Eukaryotic Linear Motif Resource (http://elm.eu.org/index.html) in the roots of the mpk6 mutant.

Accesion number	Protein name	Phosphopeptide sequences	Docking sequence	Fold change	Cell compartment
gi|15230005	regulatory particle triple-A ATPase 5A	*****MATPMVEDTS	RKGKCVVL, 102–109 [A]	unique in mpk6	proteasome
			KERFEKLGV, 194–202 [A]
gi|15232671	phospholipase D alpha 1	AAAGFPESPEAAAEA	RRPKPGGDVTI, 244–254 [A]	unique in mpk6	plasma membrane79, cytoplasm, nucleus (prediction)
			RPKPGGDVTI, 245–254 [A]
			KKKASEGVRV, 259–268 [A]
			KKASEGVRV, 260–268 [A]
			KLRDLSDIII, 439–448 [A]
			RRAKDFIYV, 511–519 [A]
			RAKDFIYV, 512–519 [A]
			KGEKFRVYVVV, 561–571 [A]
			KFRVYVVV, 564–571 [A]
gi|15232776	cell division control protein 48-A	*****MSTPAESSDS	RKKSPNRLVV, 23–32 [A]	unique in mpk6	nucleus, cytoplasm80
			KKSPNRLVV, 24–32 [A]
			KVVRSNLRVRL, 90–100 [A]
			RVRLGDVISV, 97–106 [A]
			RPVRKGDLFL, 148–157 [A]
			RKGDLFL, 151–157 [A]
			KSRAHVIV, 339–346 [A]
			RRFGRFDREIDI, 361–372 [A]
			RFGRFDREIDI, 362–372 [A]
			RFDREIDIGV, 365–374 [A]
			KNMKLAEDVDL, 389–399 [A]
			REKMDVIDL, 427–435 [A]
			RPGRLDQLIYI, 639–649 [A]
gi|15233111	cysteine synthase C1	AYDLLDSTPDAFMCQ	KRDASLLI, 50–57 [A]	unique in mpk6	plastids, cytosol, mitochondria81
			KSKNPNVKI, 238–246 [A]
			KGKLIVTI, 331–338 [A]
gi|334186086	ketol-acid reductoisomerase	APSLSCPSPSSSSKT	KKEKVSL, 88–94 [A]	3.44	plastid (prediction)
		GWSVALGSPFTFATT
gi|334185190	heat shock protein 70-3	DLLLLDVTPLSLGLE	RARFEELNI, 305–313 [A]	3.15	nucleus, cytoplasm30
			RIPKVQQLLV, 348–357 [A]
gi|145324054	arabinogalactan protein 31	EVNHKTQTPSLAPAP	KFNRSLVAV, 180–188 [A]	0.64	cel wall72
		HPHPPAKSPVKPPVK	KKLGKSTVVV, 285–294 [A]
		PPVKAPVSPPAKPPV	KLGKSTVVV, 286–294 [A]
		PPVKPPVSPPAKPPV
		APVKPPVSPPTKPPV
		PPTKPPVTPPVYPPK
gi|15221156	pyrophosphate–fructose-6-phosphate 1-phosphotransferase subunit beta 1	RDLTAVGSPENAPAK	KKAMVEL, 512–518 [A]	0.08	plastid, cytoplasm (prediction)

Abundance of abiotic stress related proteins in mpk4 and mpk6 mutants

Considerable differences were found between the mutants in proteins annotated by GO “response to stress” (Fig. 2A). Therefore we focused in detail on GO annotations related to response to external stimuli in the differential proteomes of the mutants (Fig. 4A). Some GO annotation categories such as response to starvation, zinc ion and nutrient levels were not detected in the mpk6 mutant but they were present in the mpk4 mutant. Further, significantly more differentially-regulated proteins in the mpk4 mutant, as compared to the mpk6 mutant, were annotated for response to cadmium, cytokinin, misfolded protein, auxin and hydrogen peroxide. These analyses suggested that related physiological responses might be more altered in the mpk4 mutant in comparison to the mpk6 mutant.

Figure 4.

Comparison of gene ontology categories related to (A) response to external stimuli (6^th level of ontology), (B) single-multicellular organism process (5^th level of ontology) and (C) cellular component organization (4^th level of ontology) between mpk4 and mpk6 mutants.

Both MPK4 and MPK6 are important signalling proteins during oxidative stress, albeit being activated by different signalling modules⁸,⁹. Roots of the mpk4 mutant showed decreased abundance of 7 proteins (monodehydroascorbate reductase, L-ascorbate peroxidase 1, L-ascorbate peroxidase S, peroxidase family protein, glutathione S-transferase phi 8, nucleoside diphosphate kinase 1 and peroxidase 27) involved in oxidative stress response and antioxidant defense as compared to the Col-0, while other 4 proteins (glutathione S-transferase F2, glutathione S-transferase TAU 19, monodehydroascorbate reductase (NADH) and catalase 3) were upregulated. Consistently, bioinformatic analysis of interaction networks of proteins with increased abundance in the mpk4 mutant showed the downregulation of cluster composed of enzymes involved in ascorbate-glutathione cycle (Supplemental Fig. 1A). Surprisingly, in the mpk6 mutant, we detected only one antioxidant enzyme, monodehydroascorbate reductase (NADH), which is downregulated in comparison to the wild type control. These data indicate that the Arabidopsis antioxidant defense might be regulated predominantly by MPK4.

In order to monitor the activities of hydrogen peroxide decomposing enzymes in mpk4 and mpk6 roots, we carried out specific activity staining of peroxidases and catalase on native PAGE gels. We detected four isozymes of peroxidases in both mutants and wild type. Three isozymes (with Rf 0.02, 0.03 and 0.6) showed substantial differences between Col-0 and the mutants (Fig. 5A,B,D,E). The changes in peroxidase isozymes were similar in both mutants when compared to the wild type. To analyze activities of peroxidase isozymes under oxidative stress, we exposed the wild type, mpk4 and mpk6 seedlings to 150 μM hydrogen peroxide (in liquid 1/2 MS for 24 h). The total peroxidase activity was apparently higher in the mpk4 mutant as compared to the wild type after incubation in mock treatment (liquid 1/2 MS medium), while it was reduced in the mpk6 mutant (Fig. 5C,F). Hydrogen peroxide activated one peroxidase isozyme with Rf 0.09 in Col-0 control while the other isozymes remained unaffected. On the other hand, the overall peroxidase activity was increased in response to hydrogen peroxide in both mutants (especially isozymes with Rf 0.02 and 0.03 in Fig. 5C,F).

Figure 5. Analyses of peroxidase and catalase enzymatic activities in the mpk4, mpk6 mutant and in the wild type.

(A–C) Peroxidase specific activity staining in mpk4 and mpk6 mutant roots without any treatment (A,B) or after 1 day long incubation in liquid 1/2 MS medium with or without 150 μM hydrogen peroxide (C). (D–F) Quantification of the bands optical densities in A, B and C. (G–I) Catalase specific activity staining in mpk4 and mpk6 mutant roots without any treatment (G,H) or after 1 day long incubation in liquid 1/2 MS medium without or with 150 μM hydrogen peroxide (I). (J–L) Quantification of the bands optical densities in G, H and I. (M–O) Respective loading control represented by protein visualization using stain free technology (Biorad) on SDS PAGE gels. Error bars are standard deviations calculated from 3 biological replicates.

Further, we observed significantly lower total catalase activity in the mpk4 mutant while it was slightly elevated in the mpk6 mutant when compared to wild type (Fig. 5G,H,J,K). On the other hand, catalase activity significantly increased in response to the hydrogen peroxide treatment only in the mpk4 mutant. These results show that mpk4 and mpk6 mutant roots have similar basal and hydrogen peroxide-induced peroxidase isozyme pattern compared to Col-0. In contrast to peroxidase, total catalase activity in control conditions and also in response to hydrogen peroxide showed noticeable differences between the mutants.

mpk4 and mpk6 mutants show differences in defense related proteins

MPK4 and MPK6 are activated by different MAPK cascades in response to the pathogen attack²⁰. Therefore, it is challenging to decipher whether this difference affects the composition of the defense related proteins in the mpk4 and mpk6 mutants.

We detected a canonical defense inducible protein, root specific beta-1,3-endoglucanase²¹ and DNA topoisomerase (named as NAI2)²² as upregulated in the mpk4 mutant roots, supporting well-known activation of pathogen defense in the mpk4 mutant²³. These proteins, together with JA responsive 1, form an interaction cluster showing increased abundance in the mpk4 mutant as revealed by STRING functional protein association network analysis (Supplementary Fig. S1A). Similarly, some other pathogen inducible proteins were upregulated as well. These included mitochondrial voltage-dependent anion channel²⁴, short-chain dehydrogenase reductase 3²⁵ and finally Kunitz trypsin inhibitor, which acts as antagonist of cell death triggered by phytopathogens²⁶. Strikingly, not all pathogenesis related proteins showed constitutively increased abundances in the mpk4 mutant. Thus, glutathione S-transferase 8 known as marker for early defense response²⁷ and heat stable protein 1 with antimicrobial activity²⁸ were downregulated (Supplementary Table S1). Another two downregulated proteins in the mpk4 mutant are involved in virus multiplication. The TOMV RNA binding protein binds to tomato mosaic virus genomic RNA and inhibits its multiplication²⁹. HSP70–3 interacts with Turnip mosaic virus RNA-dependent RNA polymerase (RdRp) and is a component of a replicase complex regulating RdRp function³⁰. HSP70–3 was also shown to regulate plant immune responses against bacteria in interaction with co-chaperone SGT1³¹. Therefore, MPK4 might modulate virus multiplication.

In addition to these proteins, jasmonic acid (JA) responsive protein 1 (JR1) was less abundant in the mpk4 roots (Supplementary Table S1). Since JR1 is accumulating during JA response³² it indicates downregulation of JA signalling pathway in the mpk4 mutant. On the other hand, JA-inducible lipoxygenase1 (LOX1), which is also involved in JA biosynthesis³³ was detected uniquely in the mpk6 mutant (Supplementary Table S2), suggesting upregulation of JA signalling in this mutant. Altogether, these data indicate different regulation of JA signalling by MPK6 and MPK4.

The mpk6 mutant showed also increased abundances of NADP-dependent malic enzyme 2, an enzyme important for pathogen defense³⁴ and CDC48 protein responding to pathogen infection and controlling movement of Tobacco mosaic virus³⁵. Finally, the HSP70–3, unlike in the mpk4, was upregulated in the mpk6 mutant (Supplementary Table S2).

Differential abundance of scaffold proteins RACK1 and AtRem1.3 in mpk6 and mpk4 mutants

WD-40 repeat ArcA-like protein (RACK1) was detected uniquely in the mpk6 mutant (Supplementary Table S2), while it was under the detection threshold in the mpk4 mutant. RACK1 functions as a scaffold protein linking MEKK1-MKK4/MKK5-MPK3/MKP6 cascade during immune pathway activated by pathogen-secreted proteases³⁶. In addition, RACK1 has a multiple roles in plant development and hormone responses³⁷. According to our results, MPK6 may negatively regulate the expression of RACK1 in Arabidopsis.

Another scaffold protein involved in plant-microbe interactions is remorin family protein (AtRem1.3) which showed decreased abundance in the mpk4 mutant (Supplementary Table S1). AtREM1.3 is differentially phosphorylated upon treatment with bacterial elicitors and likely plays a role as scaffold protein in plant innate immunity³⁸. Our data provide a first evidence of possible co-regulation of AtRem1.3 and MAPK signalling.

Plant developmental proteins in mpk4 and mpk6 mutants

Considerable differences between mpk4 and mpk6 mutants have been found in GO category named single-multicellular organism process (Fig. 2A). Further, we focused on downstream GO annotations within the GO hierarchy (Fig. 4B). This GO annotation contains proteins implicated mainly in developmental processes. According to this evaluation, root proteome of mpk6 contained proteins involved in megagametogenesis, pigment accumulation in tissues and tropism, while such GO annotations were not present in the root proteome of mpk4 mutant. In addition, mpk6 proteome contained a higher number of proteins involved in seed germination as compared to mpk4. On the other hand, mpk4 root proteome contained more proteins involved in seed dormancy and meristem maintenance.

Further, we manually selected differentially-regulated developmental proteins in both mutants. These proteins and their specific roles in plant development are listed in Tables 3 and 4.

Table 3. Differentially abundant proteins in the mpk4 involved in the plant development as classified according to GO annotations.

Accession	Protein name	Fold change	Developmental process
gi|18417863	14-3-3-like protein gf14 upsilon	unique in WT	root formation, chloroplast development82
gi|30682607	mRNA decapping complex vcs	unique in WT	leaf blade development83; early seedling development46
gi|30695409	acetoacetyl- thiolase 2	unique in WT	pollen tube elongation, embryogenesis84
gi|15219345	metacaspase 4	unique in WT	embryogenesis85
gi|15224838	profilin 1	0.36	root hair formation63
gi|15233538	profilin 2	0.25	root hair formation63
gi|18379240	mlp-like protein 328	0.35	bolting86
gi|15231255	tcp-1 cpn60 chaperonin family protein	0.23	plastid division87
gi|145329204	triosephosphate isomerase	0.45	transition from heterotrophic to autotrophic growth88
gi|15234781	peptidyl-prolyl cis-trans isomerase cyp1	0.54	stem elongation and shoot branching89
gi|15237054	v-type proton atpase subunit e1	0.11	embryo development90
gi|145323784	l-ascorbate peroxidase 1	0.24	embryogenesis91
gi|30686836	dehydrin erd10	0.32	seed development92
gi|145333043	adenosylhomocysteinase 1	25.25	seed and root development93
gi|18410982	s-phase kinase-associated protein 1	mpk4 unique	cell division, meiosis94,95, meristem activity96, seedling development97
gi|15233740	hsp90-like protein GRP94	4.99	shoot, floral and root meristem function45

Only proteins with published experimental evidence for their role in the plant development are listed here.

Table 4. Differentially abundant proteins in the mpk6 involved in the plant development as classified according to GO annotations.

Accession	Protein name	Fold change	Developmental process
gi|145333041	glycine-rich RNA-binding protein 2	0.136	seedling development and germination98, flower and seed development99
gi|15220770	1-aminocyclopropane-1-carboxylate oxidase 2	0.094	ethylene synthesis71, seed germination100,101
gi|15231569	aquaporin tip1-2	0.077	lateral root emergence102
gi|15228041	aquaporin tip1-1	0.053	lateral root emergence102
gi15233538	profilin 2	0.209	root hair formation63
gi|30691988	chaperone protein dnaj 3	0.136	flowering103
gi|18403295	gamma-aminobutyrate transaminase pop2	0.139	cell wall composition, root and hypocotyl development104
gi|15229522	adenosylhomocysteinase 2	1.594	seed and root development93
gi|15222075	actin 8	unique in mpk6	root hair tip growth105
gi|30683070	tubulins	5.012	helical root growth106, gravitropism107
gi|15242516	actin 7	5.805	callus formation108; seed germination and root growth109, cell division105
gi|15232671	phospholipase d alpha 1	unique in mpk6	stomatal closure110
gi|15220941	guanine nucleotide-binding protein subunit beta-like protein a	unique in mpk6	root formation, seed germination and seedling development37,111
gi|15221970	lipoxygenase 1	unique in mpk6	formation of lateral roots112
gi|15230005	regulatory particle triple-a atpase 5a	unique in mpk6	gametophyte development113, root apical meristem maintenance114
gi|15232776	cell division control protein 48-a	unique in mpk6	cell division, expansion and differentiation48
gi|15233320	aquaporin tip2-1	unique in mpk6	lateral root emergence102
gi|30691619	elongation factor 1b beta	unique in mpk6	cell wall formation and cell expansion115

Only proteins with published experimental evidence for their role in the plant development are listed here.

The cytoskeletal proteins are crucial regulators of cell division, elongation and growth. Both mutants are known for their altered organization of cortical and mitotic microtubules¹²,¹³,¹⁴. However, nothing is known about actin cytoskeleton in these mutants. Our results revealed changed abundances of several actin isoforms and actin binding proteins important for actin organization and dynamics in both mpk4 and mpk6 mutants. Thus, the mpk4 mutant showed lower levels of profilin isoforms 1 and 2 (PRF1 and PRF2) when compared to the Col-0 wild type (Table 3). This was proven also by immunoblotting analysis using antibody recognizing both PRF1 and PRF2 (Fig. 6A,B). Profilin regulates actin polymerization while Arabidopsis prf1 mutant plants display increased number of root hairs, a phenotype similar to the mpk4³⁹. Dehydrin ERD10 has actin stabilizing activity⁴⁰ and it was downregulated in the roots of the mpk4 mutant. Moreover, annexin 1 showed more than four-fold higher abundance in this mutant. Annexins are stress induced proteins involved in actin filament bundling and in vesicular trafficking in plants⁴¹. Therefore, changed abundances of above-mentioned actin binding proteins suggest modifications of actin cytoskeleton organization in the mpk4 mutant.

Figure 6.

Immunoblotting differential analyses of profilins (A,B), actin (C,D), alfa-tubulin (E,F) and beta-tubulin (G,H) in the mpk4 and mpk6 mutants. Graphs depict optical density quantifications of relevant proteins. Loading controls are provided beneath the relevant immunoblots. Error bars are standard deviations calculated from 3 biological replicates. Stars indicate significant difference at p ≤ 0.05 according to Student t-test.

Actin cytoskeleton appeared to be altered also in the mkp6 mutant, as indicated by the downregulation of actin 7 and profilin 2. On the other hand, actin 8 and actin-like ATPase superfamily protein were uniquely detected in the mpk6 mutant and this might contribute to the regulation of actin cytoskeleton organization in the mutant. Moreover, increased actin abundance in the mpk6 and decreased abundances of PRF1 and PRF2 were corroborated by immunoblotting using actin antibody recognizing all actin isoforms and profilin antibody recognizing PRF1 and PRF2 (Fig. 6A,B,C,D). Further, we detected increased abundance of alpha tubulin 5 and 6 in the mpk6 mutant. Consistently, immunoblotting analysis showed slight overabundance of tubulin isoforms in the mpk6 mutant, while lower abundances were encountered in the mpk4 mutant (Fig. 6E–H). This analysis showed that abundances of tubulin isoforms were differently regulated in both mutants.

Phospholipase D alpha 1 (PLDa1), an enzyme producing phosphatidic acid (PA) and its lipid derivatives, was uniquely detected only in the mpk6 mutant. PLDa1-derived PA binds to MAP65–1, enhances its activity during microtubule polymerization and bundling⁴². PA is capable also to activate MPK6⁴³. Our proteomic data point to possible negative regulation of PLDa1 abundance by MPK6 and this might contribute to the altered cortical microtubule organization in the mpk6 mutant¹². PLDa1 is a putative candidate for MPK6-dependent phosphorylation as suggested by bioinformatic analysis (Table 2).

Additionally, several other proteins directly involved in cell division, a process controlled by both MPK6 and MPK4 ¹²,¹⁴, were detected in both mutants (Tables 3 and 4). S-phase kinase-associated protein 1, subunit of an E3 ubiquitin ligase known as the SCF complex, was detected only in the mpk4 mutant (Table 3), suggesting its increased abundance. This protein might act downstream of MPK4 in the regulation of cell division and meristem activity. Next, CDC-48A which localizes at the cell division plane during cytokinesis and contributes to membrane fusion events⁴⁴ was overabundant in the mpk6 mutant (Table 4). Thus it might be involved in ectopic and disoriented cell divisions of this mutant¹².

Other proteins important for root growth and embryogenesis were predominantly downregulated in the mpk4 mutant, consistently with its firmly reduced growth. For example, HSP90-like protein GRP94 (known also as SHEPERD) was five-fold upregulated (Table 3). This protein is important for CLAVATA receptor kinase folding and shd mutants show root apical meristem phenotype with disorganized columella cells and altered organization of initials and central cells⁴⁵.

Proteomics revealed differences in cellular compartment morphogenesis, endopeptidase activity and RNA processing between mpk4 and mpk6 mutants

Another GO annotation showing significant differences between mpk4 and mpk6 mutants in terms of biological process was cellular component organization (Fig. 2A). The mpk6 mutant contained a higher number of proteins annotated to this class. Further evaluation showed that the mpk6 differential proteome contained a higher number of proteins annotated to morphogenesis of cellular components, membrane organization, external encapsulating structure organization and organelle organization (Fig. 4C). Moreover, the mpk6 differential proteome uniquely contained proteins involved in cellular component disassembly such as lipoxygenase 1 and glycine-rich RNA-binding protein 2.

In terms of molecular function, root proteome of the mpk4 mutant contained three GO annotations which were absent in the mpk6 mutant (Fig. 2B). The mpk4 mutant contained differentially regulated proteins with endopeptidase and metallopeptidase activity (Fig. 7A,B). Proteins with endopeptidase activity included proteases with different specificity depending on amino acids (Fig. 7A). This might indicate differences in proteolytic processes in the mutants.

Figure 7.

Graphs showing gene ontology categories related to (A) endopeptidase activity and (B) metalopeptidase activity in the mpk4 mutant.

Finally, proteomic analysis indicated decreased abundance of mRNA decapping complex VCS (known also as varicose) in the mpk4 mutant. This protein is the component of the decapping complex, that removes the 7-methyl-guanosine 5′-diphosphate from the 5′ end of mRNAs. Decapping allows rapid changes in gene expression and is important for many stress related and developmental processes⁴⁶. This protein contains the MAPK docking site and specific phosphorylation site in its amino acid sequence (Table 1) suggesting that it might be a putative substrate target of MPK6. Indeed, the MPK4-induced changes in VCS levels may contribute to deregulation of gene expression and resulting changed protein abundances in the mpk4 mutant. Nevertheless, this hypothesis should be experimentally tested in future studies.

Discussion

Considering versatility of MPK4 and MPK6 functions, OMICS technologies such as transcriptomics, proteomics and phosphoproteomics are powerful techniques to investigate different aspects of MAPK signalling¹⁵. They provide a wide range of information about transcripts and proteins being modified in MAPK mutants and help to identify new MAPK targets.

In this study, we aimed to decipher root proteomes of two Arabidopsis mpk4 and mpk6 mutants. Both MPK4 and MPK6 are known to be activated by oxidative stress and pathogens, however, they are likely involved in different MAPK signalling pathways⁴,⁵,⁸,⁹. Differential proteomic analysis of plants with genetically manipulated MAPKs could be useful for investigation of proteome-wide changes in such plants¹⁵.

Identification of putative MPK4 and MPK6 targets

It was previously reported that MAPK knock-out mutants show changes in the transcriptional levels of some respective MAPK phosphorylation targets⁴⁷. Using bioinformatic analysis we were able to predict several putative targets of MPK4 and MPK6 in analyzed proteomes. These were proteins of diverse localizations and functions and in some cases it corresponded well with known localization and function of MPK4 and MPK6. For example, CDC48A, which was detected only in the mpk6 mutant is localized to the cell plate during cytokinesis, similarly to MPK6⁴⁸,⁴⁹. This indicated that CDC48A might be a putative target protein which is regulated by MPK6.

One protein candidate involved in gene transcription in the mpk4 mutant is the mRNA decapping complex VCS which was detected only in the wild type plants (indicating its downregulation in the mpk4 mutant). This protein might be, at least partially, responsible for MPK4-dependent transcriptomic and proteomic changes in Arabidopsis roots. mRNA decapping complex VCS was also found by phosphoproteomic study of transgenic plants expressing GVG:FLAG-NtMEK2^DD (constitutively active after dexamethasone treatment) as a putative MAPK target⁵⁰. In this respect, two other proteins from our list, namely elongation factor EF2-like protein LOS1 and HSP 70–3 were also proposed as MAPK phosphorylation targets by protein microarray⁵¹.

Distinct proteome composition of mpk4 and mpk6 mutants

Our proteomic analysis showed that knockout MPK4 mutation affected broader range of proteins in Arabidopsis roots compared to MPK6 knockout. A similar pattern was observed also in the recent transcriptomic study, which reported 1235 genes differentially expressed in the mpk4 mutant but only 61 genes in the mpk6 as compared to the Col-0⁴⁷, although the difference in two mutant proteomes was not so pronounced as in the transcriptomic study. Nevertheless, it suggests that MPK4 has likely broader impact on the regulation of biological protein-dependent processes in Arabidopsis root when compared to MPK6. Next, GO annotation analysis showed functional categories unique either for mpk4 or mpk6 mutant differential proteomes. Proteins with peptidase activity as well as those involved in response to zinc, starvation and changing nutrient levels were affected in the mpk4, but not in the mpk6 mutant. On the other hand, proteins involved in several developmental processes and cellular component disassembles were altered only in the mpk6 mutant.

The relatively low number of differentially regulated proteins in mpk6 mutant root implies that the absence of MPK6 might be compensated by MPK3, which is partially redundant to MPK6. Previously it was reported that MPK3 activities could be increased in the mpk6 mutant in response flagellin in whole seedlings⁴⁷ or hypoxia⁵² and ethylene in leaves⁵³. However, little is known about basal abundance and activity of MPK3 in the mpk6 mutant and even less emphasis was given on MPK3 levels and activity in resting mutant roots so far. Published data refer about small overlap in transcriptomic profiles of 2 weeks old mpk3 and mpk6 mutant seedlings under control conditions⁴⁷ as well as leaf proteomes of 6 weeks old mpk3 and mpk6 mutants expressing MKK5^DD inducible by dexamethasone⁵⁴.

Considering MPK3 abundance, no differences were found in the leaves of 5 weeks old mpk6 mutant and wild type plants⁵⁵. These data indicate that MPK3 might not compensate for MPK6 deficiency in the leaves of the mpk6 mutant. In the 2 weeks old roots of the mpk6 mutant, we have found higher abundance but lower activity of MPK3. Thus, the capability of MPK3 to compensate for MPK6 deficiency might depend on tissue and developmental stage of plants.

The constitutive induction of pathogen defense is related to upregulation of ER body proteins in the mpk4 mutant

Present proteomic analysis confirmed some known information about defense responses in Arabidopsis roots mediated by MPK4 and MPK6. For example, while MPK6 is a negative regulator of JA signalling⁷, MPK4 positively affects the JA signalling pathway²³. This is fully consistent with elevated abundance of lipoxygenase 1, an enzyme crucial for JA synthesis, in the mpk6 roots as well as with decreased abundance of jasmonic acid responsive protein 1 in mpk4 roots found in our study.

MPK6 phosphorylates 1-aminocyclopropane-1-carboxylic acid synthase 2 and 6 leading to promotion of ethylene production⁶. We have found decreased abundance of 1-aminocyclopropane-1-carboxylate oxidase 2 in the mpk6 mutant, suggesting additional level of ethylene regulation by MPK6.

On the other hand, abundances of pathogen related proteins were not fully consistent with constitutive activations of pathogen related proteins in the mpk4 mutant⁴,²³. We detected several ER-resident pathogenesis related proteins as upregulated in the mutant. These proteins are involved in ER body formation⁵⁶,⁵⁷, which is an important component of plant pathogen defense²². NAI2, beta glucosidase and JAL34 were identified in the mpk4 mutant as upregulated, and they are controlled by transcription factor named NAI1. The analysis of MAPK-specific phosphorylation sites of NAI1 showed that MAPKs are most probable kinases responsible for phosphorylation of this transcription factor (Supplementary Table S9). NAI1 also contains docking motif for interaction with MAPK (Supplementary Table S10). All this implies that the constitutive activation of defense response in the mpk4 mutant might be connected to ER body formation. Moreover, MAPKs such as MPK4 might be crucial regulators of ER body formation controlled by the phosphorylation of NAI1.

Still, several proteins involved in pathogen defense were downregulated in the mpk4 mutant. This could be perhaps explained by the selective activation of the pathogen defense by MPK4, depending on leucine-rich receptors⁵⁸.

MPK4 is important for antioxidant defense

Members of MAPK cascades are important mediators of antioxidant defense in plants. This was shown for MKK5 in high-light induced oxidative stress for Cu/Zn SOD⁵⁹ and salinity induced FeSOD⁶⁰ as well as for MKK1 and MPK6 in the regulation of CAT1⁶¹. Here we report, that roots of mpk4 mutant possess increased levels of some proteins and enzymes involved in antioxidant defense. Similarly, increased levels of transcripts of antioxidant genes were obtained in transcriptomic study on seedlings⁹. Such increased antioxidant defense correlates with increased levels of ROS in the mpk4 mutant⁴. When compared to mpk4, mpk6 roots showed decreased abundance of monodehydroascorbate reductase.

Since these proteomic data revealed that the antioxidant defense might be differentially regulated by these two kinases, we aimed to examine enzymatic activities of peroxidase and catalase decomposing hydrogen peroxide. Our results showed that mainly catalase activity was significantly upregulated in the mpk4 mutant, unlikely to the mpk6 mutant. Previously we reported that constitutive upregulation of antioxidant defense caused increased tolerance of Arabidopsis anp2anp3 double mutant to the oxidative stress⁶². Similarly, our present data point to possible negative regulation of antioxidant defense by MPK4.

Proteomic analysis deciphering mpk4 and mpk6 root phenotypes

Both mutants have distinct phenotypes. The mpk4 mutant shows severely reduced growth²³ and short roots with radial expansion and strong root hair defects¹³. At the subcellular level, the mpk4 mutant has cytokinetic defects resulting from various abnormalities in mitotic spindle and phragmoplast rearrangements¹⁴. Knockout mpk6 mutant displays “no root”, “short root” or “long root” phenotypes at early seedling development and increased number of adventitious roots at the later developmental stages, as well as ectopic cell divisions and defective cell plate orientation¹²,⁶⁶.

In fact, some proteins directly involved in cell division were altered in both mutants. CDC-48A protein, a potential MPK6 phosphorylation target (see discussion above), which contributes to cell division plate formation⁴⁴, showed overabundance in the mpk6 mutant. Since MPK6 localizes in spots close to cell plate and determines the cell plane orientation¹² these data might indicate a negative regulation of CDC-48A by MPK6 during cell plate formation.

Concerning actin cytoskeleton, both mutants showed decreased abundance of profilins PRF1 and PRF2. Profilin is an actin binding protein crucial for root hair formation⁶³, a process which is severely altered in the mpk4 mutant¹³. Plant profilins regulate actin polymerization dynamics and filament initiation together with other actin-binding proteins (ABPs) such as actin depolymerizing factors (ADFs), actin-related proteins (ARPs) and formins. Depending on the equilibrium between these ABPs actin polymerization might be either promoted or inhibited⁶⁴. Since actin polymerization and dynamics is important for cell and organ elongation growth–hypothetically, less profilins and less actin might be connected with shorter roots in the mpk4 mutant while less profilin but more actin and ADF might result in dominating long-root phenotype in the mpk6 mutant.

In conclusion, present proteomic study identified several protein candidates contributing to the phenotypes and defense responses in the Arabidopsis mpk4 and mpk6 mutants, and paved the way for their functional characterization in the future studies.

Materials and Methods

Material

Seeds of Arabidopsis thaliana, wild type (ecotype Col-0), mpk4¹³ and mpk6–2¹² mutants were grown vertically on Petri dishes with 1/2 Murashige-Skoog solid culture medium⁶⁵ (pH 5.7) (16 h light/8 h dark, 22 °C). For mpk6–2 mutant, seedlings displaying “short root” and dominating “long root” phenotypes⁶⁶ were used for the analyses. Roots were collected for proteomic and biochemical analyses two weeks after germination. Proteomic analyses were performed on four independent biological replicates.

Protein extraction for proteomic analysis

Proteins were extracted as described previously⁶⁷. Roots of Arabidopsis wild type plants and mutants were homogenized in liquid nitrogen to fine powder and extracted in buffer containing 0.9 M sucrose, 0.1 M Tris-HCl, pH 8.8, 10 mM EDTA, 100 mM KCl and 0.4% v/v 2-mercaptoethanol. Total proteins were obtained from the extract using phenol extraction followed by precipitation in methanolic ammonium acetate (100 mM), and sequential purification in 80% v/v acetone, 70% v/v ethanol, and final incubation with 80% v/v acetone. The final protein pellet was dissolved in 6 M urea in 100 mM Tris-HCl, pH 6.8. Proteins were reduced and alkylated prior to trypsin digestion (1 μg of trypsin was applied to 50 μg of proteins). Digested peptides were desalted on C18 cartridges (Sep PAK, Waters, Milford, MA, USA) and vacuum dried.

Liquid Chromatography and Mass Spectrometry

Spectral data were collected using an Orbitrap LTQ Velos mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) using Xcalibur version 2.1.0 united with an UltiMate 3000 nano flow HPLC system (Dionex). Two μg of protein tryptic digest were loaded on reversed phase fused silica C18 column measuring 75 μm × 150 mm (Thermo Fisher Scientific). Peptides were separated and eluted at a constant flow rate of 0.3 μl.min-1 by a 170-minute long nonlinear gradient of acetonitrile (in 0.1% formic acid) as follows: 2–55% for 125 min, 95% for 15 min, 5% for 30 min. Peptides were detected in linear trap mass spectrometer operated in a data dependent acquisition (DDA) mode. The mass spectra were obtained in the data dependent mode, with dynamic exclusion being applied, in 18 scan events: one MS scan (m/z range: 300–1700) followed by 17 MSMS scans for the 17 most intense ions detected in the MS scan. Other critical parameters were set as follows: Normalized collision energy: 35%, AGC (automatic gain control) “on” with MSn Target 4 × 104, isolation width (m/z): 1.5, capillary temperature 170 °C, spray voltage 1.97 kV.

The .raw files were searched using the SEQUEST algorithm of the Proteome Discoverer 1.1.0 software (Thermo Fisher Scientific) with following selection of parameters: Lowest and highest charge: +1 and +3, respectively; minimum and maximum precursor mass: 300 and 6000 Da, respectively; minimum S/N ratio: 3; enzyme: trypsin; maximum missed cleavages: 2, FDR = 0.01; dynamic modifications: cysteine carbamidomethylation (+57.021), methionine oxidation (+15.995), methionine dioxidation (+31.990).

The spectral data were matched against target and decoy databases (later being created automatically by the software). The NCBI (www.ncbi.nlm.nih.gov) Arabidopsis genus taxonomy referenced protein database (67,924 entries as of November 2013) served as the target database, while its reversed copy served as a decoy database. The Proteome Discoverer results files (.msf) were uploaded to ProteoIQ 2.1 (NuSep) software for further filtering. Only proteins detected with at least three spectral counts, FDR < 1%, 95% probability and listed as “Top” proteins (defined by ProteoIQ, “Within a protein group, each and every respective peptide could be matched to the top protein”) are considered as high confidence matches and are presented in the results. The relative quantitative analysis was based on sums of precursor ion intensities of filtered peptides attributed to given proteins. The ANOVA analysis of four replicates for each biological sample was performed and p ≤ 0.05 was used to filter statistically significant results.

Bioinformatic evaluation of proteomic data

Proteins with statistically significantly changed abundances as well as those detected uniquely in one of the two analyzed samples were subjected to gene ontology annotation using Blast2Go¹⁷ software. The full sequences retrieved from NCBI database were blasted against Plants/Arabidopsis thaliana protein sequences allowing 1 BLAST Hit, followed by mapping step and annotation by using these parameters: E Value Hit filter: 1.0E-6; Annotation cut off: 55; GO weight: 5. One GO level was used for the evaluation of GO annotations according to molecular function, biological process and compartment.

Amino acid sequences of proteins with statistically significantly changed abundances as well as those detected uniquely in one of the two analyzed samples were screened for the presence of MAPK specific docking domains using Eukaryotic Linear Motif (ELM) resource⁶⁸. The identified proteins were further screened for the presence of MAPK-specific phosphorylation motif by using GPS 3.0–Kinase-specific Phosphorylation Site Prediction⁶⁹. Finally, Wolf Psort prediction tool (http://www.genscript.com/wolf-psort.html) was used for the prediction of protein subcellular localization.

Immunoblotting analysis of actin, tubulin and profilin abundances in MAPK mutants

Protein extracts obtained by phenol extraction were enriched with SDS sample buffer and 2-mercaptoethanol (5% v/v) and used for immunoblotting analysis. For SDS-PAGE, Stain Free^TM technology (Biorad, Hercules, CA, USA) was applied allowing UV-based visualization of proteins on gel and membrane to ensure equal sample loading. Identical protein amounts were loaded for each sample. Proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (GE Healthcare, Little Chalfont, United Kingdom) in a wet tank unit (Biorad, Hercules, CA, USA) at 100 V for 1.5 h. For immuno-detection of protein bands, the membrane was blocked in a mixture of 4% w/v low-fat dry milk and 4%w/v bovine serum albumin in Tris-buffered-saline (TBS, 100 mM Tris-HCl; 150 mM NaCl; pH 7.4) for 1 hour, and subsequently incubated with anti-profilin, anti-beta tubulin, anti-actin (Sigma-Aldrich, Germany), and anti-alpha tubulin (Abd Serotec, Kidlington, UK) at room temperature for 1.5 h. After it, they were incubated with a horseradish peroxidase conjugated goat anti-rabbit IgG secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA), diluted 1:5000 in TBS-T containing 1% w/v BSA at room temperature for 1.5 h. Following five washing steps in TBST, proteins were detected by incubating the membrane in Clarity Western ECL substrate (Biorad, Hercules, CA, USA). Luminescence was detected using Chemidoc MP documentation system (Biorad). Immunoblot analyses were performed in three biological replicates. The band signal density was quantified using Image Lab 4.0.1 software (Biorad, Hercules, CA, USA). The differences between the mutants and the wild type were statistically evaluated using Students t-test (n = 3).

Immunoblotting analysis of MPK3 and MPK6 activities

For the examination of MPK3 activity in Col-0 and mpk6 mutant roots, we extracted the proteins in E buffer (50 mM HEPES (pH 7.5), 75 mM NaCl, 1 mM EGTA, 1 mM MgCl₂, 1 mM NaF,10% v/v glycerol, Complete™ EDTA-free protease inhibitor and PhosSTOP™ phosphatase inhibitor cocktails (both from Roche, Basel, Switzerland). Following centrifugation, equal protein aliquots were subjected to overnight precipitation by addition of 5 volumes of 80% ice cold acetone. The resulting precipitates were dissolved in SDS sample buffer and boiled at 95 °C for 5 minutes. Immunoblotting was performed as described above, using polyclonal antibody against mammalian phosphorylated ERK1/2 (phospho-p44/42, pERK; Cell Signaling; Danvers, ME, USA). The differences between the mutants and the wild type were statistically evaluated using Students t-test (n = 3).

Analysis of enzymatic activities

For the assessment of enzymatic activities in untreated plants, we harvested the roots of the wild type Col-0 plants as well as mpk4 and mpk6 mutants grown 14 days on solid ½ MS. For the examination of enzymatic activities in response to oxidative stress treatment, 14 days old seedlings were incubated either in liquid ½ MS medium, or in the same medium supplemented with 150 μM H₂O₂ for 24 h. For preparation of native protein extract, roots were homogenized in liquid nitrogen and the homogenate was incubated with 50 mM sodium phosphate buffer (pH 7.8) containing 1 mM EDTA, 10% v/v glycerol and “Complete EDTA-free protease inhibitor cocktail (Roche, Basel, Switzerland). Following centrifugation at 13000 g at 4 °C, protein content was estimated in supernatants using Bradford assay⁷⁰. Equal protein amounts were used for further analyses. Isozymes of peroxidases and catalase were separated on 12% native PAGE at constant 10 mA/gel. For visualization of peroxidases, gels were first equilibrated in 50 mM sodium acetate (pH 5,2) buffer for 15 min, followed by incubation in 0.05% (w/v) o-dianizidine and 3% (v/v) hydrogen peroxide in 50 mM sodium acetate (pH 5.2). Catalase isozyme was visualized according to Aebi⁷¹. Gels were three times washed in distilled water for 5 minutes followed by incubation in 0.006% (v/v) hydrogen peroxide for 10 min. Catalase activity appeared as negative bands after 5 min of gel incubation in 1% (w/v) potassium ferricyanide and 1% (w/v) ferric chloride in distilled water. The band intensities were calculated using Image Lab 4.0.1 software (Biorad). Analyses were performed in biological triplicates. Statistical evaluation of data was carried out using Student’s t-test. As a loading control equal amounts of protein were resolved by SDS-PAGE as described above with the exception that precast 4–12% gradient gels (Mini-PROTEAN TGX Stain-Free Precast Gels, BioRad) were used allowing UV-induced total protein visualization with Chemidoc MP imaging system (BioRad). All experiments were carried out in biological triplicates.

Additional Information

How to cite this article: Takáč, T. et al. Comparative proteomic study of Arabidopsis mutants mpk4 and mpk6. Sci. Rep. 6, 28306; doi: 10.1038/srep28306 (2016).