Raw Nt proteomics data of Arabidopsis cell cultures acquired previously (Willems et al., 2017) were re-analyzed for comprehensive identification of Nt proteoforms. First, RAW files were converted to peak lists [Mascot Generic Format (MGF) files] using ThermoRawFileParser (Hulstaert et al., 2020). Resulting peak lists were searched with COMET (Eng et al., 2013) against a custom protein database (available at the Open Science Framework project “ajx5e” [https://osf.io/ajx5e/]) consisting of TAIR10 protein entries (such as splice forms) appended with novel proteoform sequences initiated from 15,741 Ribo-seq-called TIS not annotated in TAIR10 (Supplementary Dataset S1). It should be noted that 52 of these TIS were, however, annotated in the more recent Araport11 re-annotation (Cheng et al., 2017). Variable modifications for database searching included Nt light/heavy acetylation (+42.011 or 47.036 Da, respectively) and pyro-Glu formation from Gln (−17.026 Da). Fixed modifications were Met oxidation (+15.995 Da), Lys heavy acetylation (+47.036 Da), and Cys carbamidomethylation (+57.021 Da). For the respective datasets, Trypsin, Glu-C, Asp-N, and chymotrypsin were set as a digestive enzyme, specifying two missed cleavages for trypsin and three missed cleavages for all other enzymes. For the searches, no enzymatic digestion rules apply to the N-terminus of the searched peptides (Comet option “num_enzyme_termini = 9”) to additionally enable the identification of database non-annotated Nt peptides. Percolator output files were processed by Percolator (version 3.05.0) (The et al., 2016) using the options “–Y –trainFDR 0.05 –testFDR 0.05” and resulting Nt peptide spectral identifications (PSM q-value ≤ 0.01) were parsed. In the overview list of protein N-termini reported (Supplementary Dataset S2), Nt peptides with an identical protein start position but varying at their C-termini because of digestion by different enzymes or missed cleavages were collapsed to the longest Nt peptide variant, thereby reducing redundancy and improving the uniqueness of peptide-to-protein assignments.

pUNI51 vectors (Arabidopsis Biological Resource Center [ABRC], Columbus, OH, United States) encoding full-length isopentenyl diphosphate isomerase 2 (IPP2, AT3G02780, stock number: U22155), plant UBX domain-containing protein 1 (PUX1, AT3G27310, stock number: U82310), and NAC domain-containing protein 14 (NAC014, AT1G33060, stock number: U88651) served as templates for site-directed PCR mutagenesis (QuickChange; Stratagene, San Diego, CA, United States). For cloning and propagation of pUNI51 vectors, the Escherichia coli (E. coli) strain PIR1 (One Shot^TM PIR1 chemically competent E. coli, cat n°C101010; Invitrogen, Waltham, MA, United States) was used using standard chemical transformation protocols (i.e., pUNI51 contains a conditional origin of replication derived from the R6Kγ plasmid that enables propagation only in bacterial hosts expressing the pir gene encoding the essential replication protein p) and with selection in the presence of 50 μg/ml kanamycin. Site-directed PCR mutagenesis was performed according to the manufacturer's instructions and using the primer pairs indicated in Supplementary Table 1 to mutate the riboproteogenomics-identified ATG start codons to the Leu codon TTG. The correctness of all (mutant) cDNA insert sequences was confirmed by Sanger sequencing using the pUNI51 forward and reverse primers 5′-CTGTTGGTGTGTCTATTAAATCG-3′ (pUNI51-fwd) and 5′-TGGCTGGCAACTAGAAGGCAC-3′ (pUNI51-rev), respectively.

To inspect translation products, the mutagenized constructs served as templates for in vitro coupled transcription/translation using a rabbit reticulocyte lysate system according to the manufacturer's instructions (TnT T3 Coupled Transcription/Translation Lysate System; Promega, Madison, WI, United States). More specifically, [³⁵S] methionine-labeled translation products were generated using the TnT T3 RNA polymerase and 2 μl of [³⁵S] methionine (10 μCi/ml) per reaction. After 1 h of incubation at 30°C, and to stop the translation reaction, 4 μl of the reaction mixture (50 μl reaction mixture in total) was diluted 12.5-fold in 50 mM Tris (pH 8) and NuPAGE^® LDS Sample Buffer (Invitrogen, Waltham, MA, United States), and the samples were heated for 10 min at 70°C. The samples were separated on 4–12% or 12% NuPAGE^® Bis-Tris gradient gels (1 mm × 12 well; Invitrogen, Waltham, MA, United States) using MOPS Buffer. Subsequently, proteins were transferred into a PVDF membrane, air-dried, and exposed to a film suitable for radiographic detection (ECL Hyperfilms; Amersham Biosciences, Buckinghamshire, United Kingdom).

The stand-alone version of TargetP 2.0 (Almagro Armenteros et al., 2019) was used to predict cleavage sites of Nt sorting signals. iceLogo (Colaert et al., 2009) sequence motifs were generated on the online webserver (https://iomics.ugent.be/icelogoserver) using the precompiled Swiss-Prot composition of Arabidopsis as reference set. Nt peptide data stored in the Plant PTM Viewer (Willems et al., 2019) and NTerDB (https://n-terdb.i2bc.paris-saclay.fr/) were cross-referenced to the Ribo-seq-called TIS data. Additionally, proteins were appointed a subcellular location based on SUBA4 consensus locations (Hooper et al., 2017).

All mass spectrometry proteomics data and search results of Arabidopsis cell cultures have previously been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al., 2019) partner repository with the dataset identifier PXD004896 (Willems et al., 2017). Ribo-seq sequencing data have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession GSE88790. The PROTEOFORMER Ribo-seq MySQL database and its derived protein FASTA database and the mass spectrometry proteomics data corresponding to the Nt proteomics data of chloroplast and stromal enriched fractions of Arabidopsis leaves are all made available under the Open Science Framework project “ajx5e” (https://osf.io/ajx5e/).

We previously used Ribo-seq and Nt proteomics data in a complementary fashion for delineating unannotated protein-coding ORFs in intergenic regions of Arabidopsis (Willems et al., 2017). Using the translation inhibitors lactimidomycin (LTM) and cycloheximide (CHX), ribosomes were halted to determine ribosome occupancy during translation initiation and elongation, respectively, thereby providing complementary evidence for unannotated protein start sites. Here, we re-purposed the Ribo-seq data to discover alternative TIS (aTIS) within annotated protein-coding regions located downstream from database-annotated protein start sites (dbTIS) indicative of translation initiation events potentially giving rise to the expression of Nt proteoform pairs. As previously shown, a normalized distribution of ribosome footprints after CHX and LTM inhibition display strong signals at TAIR10-annotated starts and a triplet periodicity indicative of translation (Figure 1A). For instance, a strong LTM peak clearly and uniquely delineates the dbTIS of Arabidopsis CYTOCHROME C-1, while CHX coverage spans all three exonic regions (Figure 1B). Next, we used the high-quality CHX and LTM data (Figure 1A) for TIS calling in the Arabidopsis genome using PROTEOFORMER (Crappe et al., 2015) (see section Materials and methods). This resulted in the automated Ribo-seq-based calling of 29,013 TIS (Supplementary Dataset S1), including 13,324 TIS (46%) mappings to TAIR10/Araport11 dbTIS of 13,069 genes (Figure 1C). Besides the 71 TIS located in 3′ leader sequences (so called 3′ UTRs), 7,572 called TIS reside in the 5′ leader sequences (5′ UTRs) of 4,486 genes, corroborating the widespread occurrence of upstream ORF (uORF) translation events in Arabidopsis (Von Arnim et al., 2014; Niu et al., 2020). Interestingly, another 7,653 TIS were located in the CDSs of 2,818 genes, including possible downstream TIS (dTIS) that could potentially give rise to alternative, truncated Nt proteoforms in the case of in-frame dTIS. Considering an additional 393 TIS peaks residing outside of annotated genes, more than half of the called TIS (15,689 TIS) hint at database non-annotated TIS or alternative TIS (aTIS) (Figure 1C). These aTIS show substantial initiation at non-AUG, near cognate start codons (i.e., codons differing from AUG by a single nt). More specifically, besides 20.8% of AUG aTIS, initiation at near-cognate start codons, such as CUG (14.1%), AUC (12.6%), AUU (11.6%), and UUG (9.74%) was prevalent (Figure 1D). While a similar start codon usage of non-annotated TIS can be observed for the 5′ and 3′ leader sequences as well as the CDSs of annotated genes, a significant larger proportion of 169 AUG start codons (43%, χ² test p < 0.001) was apparent among the 393 TIS identified in so-called intergenic regions (Supplementary Figure 1). This category of TIS may reflect the protein-coding potential of pseudogenes or transposable elements, besides un-annotated TIS upstream of dbTIS giving rise to Nt extended proteoforms or, alternatively, the discovery of novel gene products as reported previously (Willems et al., 2017). Taken together, Ribo-seq TIS calling mapped a wealth of unannotated (54%), besides annotated, TIS (15,689 vs. 13,324 TIS, respectively), suggesting a widespread unexplored TIS landscape in Arabidopsis (Figures 1C,D).

To gain complementary evidence of protein synthesis supporting Ribo-seq-called TIS, we re-analyzed the positional proteomics of Arabidopsis cell suspension culture data previously generated (Willems et al., 2017). More precisely, these data concern Nt proteomics datasets of proteomes digested with either trypsin, chymotrypsin, or the endoproteinases Glu-C and Asp-N shown to increase the overall Nt proteome coverage of Arabidopsis. In addition to TAIR10 protein sequences, we supplemented our search database with (non-TAIR10) protein sequences corresponding to all in silico translations matching Ribo-seq-called TIS. In total, we identified Nt peptides matching 6,493 unique protein N-termini supported by at least two peptide-to-spectrum matches (PSMs; q ≤ 0.01) (Supplementary Dataset S2). Of these, 2,041 mapped N-termini (31.4%) corresponded to TAIR10 dbTIS (protein start position 1 or 2). Except for four protein N-termini (20 PSMs, 0.1%), all identified protein N-termini matching dbTIS were compliant with N-terminal methionine excision (NME) enzymatic specificity (Bienvenut et al., 2012) (Figure 2A, for number of PSMs see Supplementary Figure 2). Moreover, based on the number of (in vitro) Nt acetylated (NTA) peptides, NTA patterns corresponded with the known specificities and NTA efficiencies of major NAT (NatA, NatB, and NatC/E/F) enzymatic activities in eukaryotes (Ree et al., 2018) (Figure 2A). Taken together, the identified Nt peptides matching dbTIS abide known co-translational enzymatic hallmarks at protein N-termini, which highlights the high quality of the obtained N-terminal proteomic data.

Next to Nt peptides matching annotated TAIR10 N-termini (dbTIS), complementary Nt peptide was evidenced for 81 Ribo-seq-called aTIS (Supplementary Dataset S2; Figure 2B). The majority of these aTIS (i.e., 72 out of 81 aTIS [88.9%]), corresponded to in-frame dTIS in TAIR10 CDS, thus pointing to Nt-truncated proteoforms (Figure 2B). In addition, four identified N-termini matched an aTIS residing in the 5′ UTR reflecting uORF or 5′ leader translation. Lastly, five N-termini matched TIS of intergenic ORFs within the so-called pseudogenes described previously (Willems et al., 2017). Hence, in addition to Ribo-seq evidence, Nt proteomics indicates the common translation of Nt-truncated proteoforms initiated from dTIS within TAIR10 CDSs. Of the 72 N-termini identified pointing to dTIS, four N-termini were, however, identified (16 PSMs, 2.9%) that did not comply with the enzymatic rules of NME, as they exposed large, bulky residues (Ile, Lys, Glu, or Gln). In addition, these four N-termini were initiated from near-cognate start codons and, hence, are more likely to represent possible neo-N-termini and, thus, false positive TIS matches and were not considered further. All the 68 other N-termini identified matching Ribo-seq-called dTIS (529 PSMs, 97.1%) were NME-compliant and showed NTA patterns consistent with the NAT enzymatic rules (Figure 2C). Similar proportions of (modified) N-termini were observed, except for relatively higher proportions of Ac-Thr (7.7% dTIS vs. 2.1% dbTIS) and unmodified Ser N-termini (2.6% dTIS vs. 0.2% dbTIS) (χ² test adjusted p < 0.05). Taken together, these Nt peptide identifications and their modification status provide unambiguous complementary proteomics and Ribo-seq proof of the translation of Nt-truncated proteoforms from dTIS for the 68 genes (Table 1). Next to two AGG, two CTG, and a single ATA, ACG, and TTG near-cognate dTIS, AUG was expectedly the most frequent start codon (61/68) among the dTIS (Supplementary Dataset S2). Next to the matching Nt proteomics Arabidopsis cell culture data, we cross-referenced publicly available Nt peptide data of Arabidopsis stored in the Plant PTM Viewer (Willems et al., 2019) and NTerDB (https://n-terdb.i2bc.paris-saclay.fr) with the Ribo-seq-called dTIS identified in this study. Overall, Nt peptide evidence was also found for 42 out of 68 dTIS-indicative N-termini in various Arabidopsis studies, and 34 of these were identified as NTA (Supplementary Dataset S2). An additional 20 Nt peptides identified in various plant tissues matched the Ribo-seq-called dTIS not identified in our Nt COFRADIC analyses (Supplementary Dataset S3). Moreover, all of them were initiated from AUG start codons, and all abided to the NME and NTA enzymatic rules, with 17/20 being NTA except for a Val-starting N-terminus known to be frequently unmodified or partially NTA, and an MP- and MK-starting N-terminus representative of a substrate refractory toward NTA (Goetze et al., 2009) and a typical partial NTA NatC/NatF substrate (Figure 2A). Hence, taking additional advantage of publicly available Nt proteomics data, we were able to obtain univocal riboproteogenomics proof of 88 dTIS-initiated Nt truncated proteoforms, many of which were expressed in a variety of Arabidopsis tissues.

To further support alternative translation initiation at in-frame dTIS as a source for the generation of N-terminal proteoform(s) (pairs), we tested if the mutagenesis of annotated TAIR10 TIS (dbTIS) and/or Ribo-seq-called dTIS would result in altered proteoform expression by in vitro coupled transcription and translation (TnT). First, we tested ISOPENTENYL DIPHOSPHATE ISOMERASE2 (IPP2), for which Ribo-seq LTM evidence suggests translation initiation from both the annotated dbTIS and dTIS possibly resulting in the expression of an IPP2 proteoform pair (Figure 3A, green and orange arrowheads, respectively). Alternative translation initiation from a downstream AUG start codon (Met53) indeed resulted in the translation of a 52 AA Nt-truncated proteoform besides the translation of the full-length annotated protein (Figure 3A). In addition, protein synthesis from the identified dTIS was evidenced by three unique NTA peptides (18 PSMs), “TDTKDAGMDAVQR” (trypsin-digested), “TDTKDAGMDAVQRRL” (chymotrypsin-digested), and “TDTKDAGMDAVQRRLFE” (Glu-C-digested), and tryptic Nt peptide evidence previously reported in four Arabidopsis studies (Venne et al., 2015; Zhang et al., 2015, 2018; Mielke et al., 2021). Mutation of the dTIS corresponding to the N-terminus starting at AA position 52 and the dbTIS of IPP2 (resulting in M1 or M52>L mutation) resulted in the exclusive expression of the Nt proteoform translated from the non-mutated TIS (Figure 3A). Second, TnT analysis of the UBX DOMAIN-CONTAINING PROTEIN1 (PUX1) was performed, for which a dTIS matching AA position 22 was identified in our riboproteogenomics analysis (Figure 3B). Also here, three unique NTA peptides matching the dTIS were identified (19 PSMs), further supported by Nt peptide evidence obtained in three other Arabidopsis studies (Bienvenut et al., 2012; Zhang et al., 2015, 2018). Furthermore, mutation of the M1 or M22 encoding codons led to exclusive Nt proteoform expression initiated at the dTIS (26 kDa) or dbTIS (29 kDa), respectively. Lastly, we mutated an identified dTIS matching AA position 12 in the transcription factor NAC DOMAIN-CONTAINING PROTEIN14 (NAC014) (Figure 3C) supported by a single NTA peptide (2 PSMs). Next to observing Nt proteoform expression starting at M1 and M13 (74 and 72 kDa, respectively), multiple lower weight proteoforms from possible cryptic downstream TIS increased in abundance upon mutation of TIS-encoding M1 and M13. Taken together, the co-expression of at least two Nt proteoforms (proteoform pair) could be observed for IPP2, PUX1, and NAC014 in vivo (Ribo-seq and Nt peptide evidence) and in vitro (TnT). Besides, mutations of TIS further confirm downstream translation initiation, likely via leaky ribosome scanning, giving rise to the expression of Nt-truncated proteoforms and, thus, the possible (co-)expression of proteoform pairs from a single CDS.

Of the 2,135 identified NTA N-termini (at least one PSM), 426 (20%) N-termini were not directly indicative of translation initiation, as they did not match dbTIS or Ribo-seq-called dTIS (Supplementary Dataset 2). Plotting the number of PSMs of identified NTA N-termini according to their start position in the protein sequence clearly shows, besides dbTIS density (protein position 1/2), increased density in the Nt protein region (e.g., position < 100) (Supplementary Figure 3A). Exceptionally, in photosynthetic species, after processing of their sorting signals, chloroplastic localized proteins can be post-translationally NTA by GCN5-related N-acetyltransferase (GNAT) domain-containing proteins, which show diverse NAT substrate specificities (Rowland et al., 2015; Bienvenut et al., 2020). In accordance, the protein position of 185 of the 426 non-TIS-called NTA N-termini (43%) deviated <5 AA residues from the TargetP 2.0 (Almagro Armenteros et al., 2019). SPP predicted cleavage site corresponding to the removal of the cTP (Figure 4A). However, 149 out of 185 of these NTA peptides also had corresponding Nt-free peptide identifications, which, in the case of Ala and Ser N-termini, were more abundant when considering PSMs opposed to Val and Thr N-termini (Figure 4B, left). Furthermore, 332 additional N-termini matching SPP cleavages (and with 513 N-termini matching cTP cleavages identified in total) solely were identified as Nt-free (Supplementary Dataset S2; Figure 4B, right). Aside from predicted cTP, the start position of another 25 NTA N-termini deviated <5 AA residues from predicted mitochondrial transit peptide (mTP) cleavage sites, while none matched signal peptide and luminal transit peptide predicted cleavages (Supplementary Figure 3B). As no NTA enzymatic activity is currently known in mitochondria (Giglione and Meinnel, 2021), these may hint at ambiguous targeting peptides representing cTP instead of mTP sequences or, alternatively, point to dual protein localization explained by the expression of distinct Nt-modified proteoforms (e.g., the NTA and Nt-free proteoform variants localizing in the chloroplasts and mitochondria, respectively).

To relate our Arabidopsis cell culture findings with in vivo degrees of NTA and to confirm the location of NTA neo-N-termini indicative of post-translational cTP-processing in the chloroplast, we additionally performed Nt COFRADIC analyses on proteomes of enriched chloroplasts and corresponding stromal protein fractions from Arabidopsis leaves. Both fractions were processed in two different ways. First, as performed for cell cultures (Willems et al., 2017), heavy ¹³C₂D₃-acetate labeling was performed in vitro, which enables quantitative determination of the degree of in vivo NTA, equaling the ratio of light over heavy NTA peptide precursor intensity (Δ 5 Da) (Van Damme et al., 2011, 2013). Second, another Nt COFRADIC analysis was performed where the invitro¹³C₂D₃-acetate labeling step was omitted. This leaves non-NTA peptide N-termini unmodified, thereby solely enriching in vivo NTA N-termini. In total, 311 natural NTA peptides were identified in the chloroplast and 848 in stroma, the latter enriched for natural NTA peptides by serving a sink of cTP-processed NTA neo-N-termini, with 127 NTA peptides common to both (Supplementary Dataset S4). A total of 297 (35%) NTA peptides correspond to cTP cleavage processing events (within the P5–P5′ region of predicted cTP cleavage sites). Of these, 103 NTA peptides were previously identified in Arabidopsis cell cultures and, thus, corroborate post-translational NTA in the chloroplast as additional NTA source in the riboproteogenomics analysis. In addition, 26 NTA peptides matched mTP cleavage sites, of which the corresponding Nt proteoforms are likely dual localized or incorrectly predicted by TargetP 2.0 as indicated before. As trimming of neo-N-termini by aminopeptidases is prevalent in the chloroplast (Rowland et al., 2015), we used the likely primary SPP cleavage site by selecting the most upstream NTA neo-N-terminus with a start position in the P5–P5′ region of predicted cTP/mTP cleavages of proteins assigned a chloroplastic consensus localization according to the subcellular localization database for Arabidopsis proteins SUBA4 (Hooper et al., 2017). This resulted in 194 NTA representative neo-N-termini, with frequent occurrences of Cys, Ala, Arg, and Met in the P1 position and Ala, Ser, and Val in P1′ (Figure 4C). While Cys, Ala, and Met were observed earlier at P1 of neo-N-termini generated upon cTP cleavage in Arabidopsis, Arg and other basic residue occupancies were suggested earlier by considering a high-confident subset of recombinant primary SPP cleavage sites (Rowland et al., 2015). In addition, enrichment of Ala and Cys at P1, as well as Val and Ile at P3, matches the characteristic SPP P3–P1′ cleavage motif reported before, (V/I)-X-(A/C)↓A (Gavel and Von Heijne, 1990). Despite SPP cleavage site hallmarks, it is, however, likely that aminopeptidase activity still complicates delineation of primary SPP cleavage motifs (Figure 4C).

Next to SPP cleavage events, and while rationally fewer than in global N-terminome analyses, 503 NTA peptides matching annotated protein N-termini (position 1 or 2) were also identified. Besides the 17 annotated protein N-termini of chloroplast-encoded genes, the remainder are from cytosolic origin and were likely identified as impurities in the chloroplast isolation procedure as also evidenced by their general greater cellular abundance. For all NTA peptides, the degrees of NTA were calculated (Supplementary Dataset S4) and plotted for dbTIS-indicative protein N-termini (position 1 or 2). The degree of NTA largely agrees with the earlier observed PSM counts of protein N-termini in cell cultures (Figure 2A) and known enzymatic NAT efficiencies (Figure 4D, left), with high degrees of NTA of Ser and Ala N-termini (average 94.6% and 80.3%, respectively) [NatA substrates (Linster et al., 2015)] and MD/ME/MN/MQ N-termini (NatB) (Van Damme et al., 2012). For Thr N-termini, there seems to be a bimodal distribution of high and low NTA N-termini, while the degree of NTA is generally low for Gly (21.4% NTA on average), Val (1.4%) (NatA) and MK (5.4%) N-termini (NatF) (Van Damme et al., 2011), and, as expected, Pro N-termini were essentially Nt-free (Goetze et al., 2009). Next, we inspected the degree of post-translational NTA of all unique stromal neo-N-termini. Sufficient NTA peptide identifications were evident for Ala, Ser, Val, and Thr neo-N-termini, clearly showing a differing degree of NTA compared to annotated protein N-termini. While the NTA degree was generally lower in case of Ser and Ala (36.3 and 34.9%, respectively), it was higher for Val (60.3%) (Figure 4D, right), and the distribution of Thr was rather centered as compared to the bimodal distribution observed for annotated protein N-termini. Similar co- and post-translational NTA degrees of (neo-)N-termini were evident in both the chloroplast fractions analyzed, albeit with less observations in case of whole chloroplasts (Supplementary Figure 4). In addition, this quantitative NTA analysis is in line with observed NTA and Nt-free PSM identifications in cell cultures (Figure 4B, left). The deviating NTA patterns of stromal neo-N-termini vs. translation-indicative Ala, Ser, Thr, and Val N-termini likely reflect differing enzymatic specificities and efficiencies of the responsible chloroplast GNAT as compared to cytosolic NatA; the latter NAT is responsible for the co-translational NTA of Ala, Ser, Thr, and Val N-termini. Taking together, post-translational NTA in the chloroplast forms a major source of NTA peptides in plants and shows different enzymatic preferences, and thus NTA patterns, compared to co-translational NTA. These results are in line with previous reports (Rowland et al., 2015; Bienvenut et al., 2020).

Thus far, we have shown that NTA peptides matching internal protein start positions may either result from protein synthesis at dTIS (because of co-translational NTA) (Figure 2C) or, alternatively, represent mature neo-N-termini of nuclear-encoded chloroplastic proteins that were post-translational NTA (Figures 4B,D). However, in some cases, it is complicated to attribute NTA peptides to either process. For instance, Met residues are well-represented in the P1 position of stromal NTA neo-N-termini (30/150 sites [20%]), therefore, representing possible TIS (Figure 4C) and 34 out of 68 (50%) proteins with a dTIS called and matching Nt peptide evidence had predicted TargetP 2.0 cleavage sites (17 cTP, 16 mTP, and single luTP). This is significantly higher as expected, since cTP/mTP/luTP are predicted for approximately 10% of the TAIR10 proteome by TargetP 2.0. Moreover, for the majority of these proteins (26 or 76%), the predicted TargetP 2.0 cleavage site deviated <10 residues from the corresponding dTIS called (Supplementary Dataset S2). For instance, the Ribo-seq-called dTIS of IPP2 (Figure 3A) exactly matched a predicted cTP cleavage with M52 at P1. Hence, NTA either occurred co- or posttranslational in the cytosol or chloroplast stroma, respectively, or both scenarios could have occurred. In line, for its close homolog IPP1, NTA peptides were identified in cell cultures starting at Ala60 and Ser53, corresponding respectively to a Ribo-seq-called dTIS at Met59 and a predicted cTP cleavage site at position 54 (Figure 5A, top). Moreover, (neo-)N-termini matching the by Ribo-seq-identified dTIS (Thr60 starting N-terminus) or cTP cleavages (Ala53, Ser55, and Ala56 starting N-termini) were found in the chloroplast leaf extracts. In chloroplast stroma, an NTA degree of 80% was identified for Ala53, whereas lower percentages of NTA were observed for Ser55 (28%) and Ala56 (27%) (Figure 5A, bottom; Supplementary Dataset S4). Hence, the presumably primary SPP-exposed Ala53 neo-N-terminus shows a higher NTA degree compared to its downstream, aminopeptidase-processed proteoforms, possibly due to the kinetics of ensuing Nt modifications. In addition, the highest degree of NTA was observed in the case of the Thr60 (91%) Nt peptide raised by alternative translation initiation at the dTIS corresponding to Met59 and subsequent co-translational removal of this iMet by the action of MetAPs. Confirming our results, for both IPP homologous genes, a long and short transcript has previously been discovered in Arabidopsis because of alternative transcription start sites, encoding full-length proteins targeted to mitochondria or chloroplasts, or Nt truncated cytosolic proteoforms (Phillips et al., 2008). In addition, Nt peptides supporting dTIS as well as organellar import processing were also found for ω-amidase (AT5G12040.1). In fact, an alternative start codon corresponding to Met63 was suggested earlier as a possible dTIS (Zhang and Marsolais, 2014) and here verified in cell cultures by Ribo-seq and NTA peptide evidence matching the iMet-processed N-terminus at Ala64 (Figure 5B, left). In addition, Ala64 was predicted as a cTP cleavage site and was found to be partial NTA (76%) in stromal fractions (Figure 5B, right). In addition, we identified an NTA peptide upstream at Ser61 in the stromal fractions, likely representing the primary exposed SPP cleavage site that was NTA to a lower extent (37%). This NTA percentage is in line with the ratio of Nt-free (16) over NTA PSMs (9) for the neo-N-terminus identified in cell cultures (Figure 5B, left). Taken together, downstream in-frame translation start sites are frequently found for proteins targeted to the chloroplast and/or mitochondrion and dTIS thereby function as a (regulatory) mechanism for acquiring functional protein copies in the cytosol (i.e., translated proteoforms lacking signal sequences of their full-length counterparts) as well as in organelles (i.e., translated proteoforms encompassing signal sequences). Moreover, the degree of NTA can provide an additional cue to distinguish Nt proteoforms originating from alternative translation or organellar processing.