Introduction

For eukaryotic cells to maintain genomic stability, the physical ends of their linear chromosomes require protection from being recognized as double-strand breaks and subsequent illegitimate end-to-end fusions. Telomeres consist of highly repetitive DNA sequences bound by dedicated proteins that provide such protection by capping chromosomes and promoting successful DNA replication through the DNA terminus to facilitate cell proliferation (Bonnell et al. 2021). The length of telomeric DNA tracts is established and maintained at a species-specific set point, though it varies widely among different organisms and even among populations of the same species (Shakirov et al. 2022). While many genes are known to control telomere length homeostasis, our current understanding of the genetic and molecular mechanisms that establish the initial telomere length set point remains limited. Genome-wide studies in multiple species implicated hundreds of genes in telomere length regulation (Askree et al. 2004; Codd et al. 2013; Li et al. 2020; Choi et al. 2021). Recent evidence also indicated that some of these genes function in telomere length control indirectly and suggested that functional crosstalk occurs between mechanisms that modulate telomere length homeostasis, chromatin modification, rRNA biogenesis, cell proliferation and other physiological and developmental pathways (Xie and Shippen 2018; Wang et al. 2023; Shakirov et al. 2022).

Using QTL mapping and transgenic manipulations in the model plant Arabidopsis thaliana, we recently identified OLI2/NOP2A, which encodes a ribosomal RNA methyltransferase, as a novel positive regulator of telomere length (Abdulkina et al. 2019). Plants deficient in the OLI2/NOP2A lose up to 27% of their telomeric DNA. By the third mutant generation (G3) these plants establish a new shorter telomere length set point that is stable for at least three subsequent generations. The human homolog of Arabidopsis OLI2/NOP2A, dubbed NOP2 (Bourgeois et al. 2015), functions as a S-adenosyl-L-methionine dependent methyltransferase in ribosome biogenesis, cell proliferation, and cancer progression (Freeman et al. 1988; Ranzani et al. 2013). Notably, the A. thaliana OLI2/NOP2A was previously identified in a forward genetic screen for genes involved in coordination of cell proliferation and cell expansion (Fujikura et al. 2009). Besides OLI2/NOP2A, several other OLIGOCELLULA (OLI) genes were identified, including OLI5/RPL5A and OLI7/RPL5B, which encode the Arabidopsis paralogs of ribosomal protein L5 (Pinon et al. 2008). RPL5 binds specifically to 5 S rRNA and serves as an essential structural component of the large 60 S ribosomal subunit (Mathieu et al. 2003).

Coordination of cell division and cell proliferation is a critical developmental process. In Arabidopsis, mutations in a group of genes, including ANGUSTIFOLIA3 (AN3), result in a decreased number of leaf cells and lead to a significant postmitotic increase in the size of remaining cells to maintain nearly normal leaf area, a phenomenon known as “compensation” (Nozaki et al. 2020). For example, in an3 mutant leaves, the number of subepidermal cells drops to ~ 30% of wild type, but these plants compensate for lost leaf area by increasing the size of the remaining cells to 150% of those in the wild type (Horiguchi et al. 2006). Thus, cell proliferation and cell expansion are highly coordinated events necessary to allow proper organ development (Horiguchi and Tsukaya 2011). AN3, also known as Arabidopsis thaliana GRF-INTERACTING FACTOR1 (AtGIF1), is a transcriptional co-activator that promotes cell proliferation. Its homologs are also involved in ribosomal biogenesis in eudicots (Vercruyssen et al. 2014) and modulate transformation efficiencies in monocots, possibly through roles in stem cell maintenance (Lee and Wang 2023). AN3 encodes a homolog of the human transcription co-activator known as synovial sarcoma translocation protein (SYT) (Brett et al. 1997; Horiguchi et al. 2006). Both SYT and AN3 coactivators interact with the chromatin remodeling SWI/SNF complexes (i.e., BRAHMA) to stimulate expression of a suite of cell proliferation and cell cycle genes (Vercruyssen et al. 2014; Yu et al. 2019).

Not all mutations that affect cell proliferation in leaves trigger compensation. Plants bearing single mutations in any one of the three OLIGOCELLULA genes (OLI2, OLI5 or OLI7) do not exhibit compensation, despite showing up to 50% reduction in leaf cell numbers. However, combining any of these mutations with an3 leads to markedly enhanced compensation, implying that compensation is regulated by a threshold-dependent mechanism (Fujikura et al. 2009). Previous findings implicating OLI2, OLI5 and OLI7 in ribosome biogenesis further indicate that compensation and organ size more generally may be under dual regulation by AN3 and ribosome-related processes.

We previously reported that several OLI genes affecting organ size control are involved in telomere biology and serve as positive regulators of telomere length (Abdulkina et al. 2019). Since OLI2, OLI5 and OLI7 control cell proliferation through their interaction with the AN3-dependent pathway, here we investigate the role of the AN3 gene in telomere biology. Using a multigenerational analysis, we report that A. thaliana an3 mutants display telomeres shorter than wild type. Moreover, double an3 oli2 mutants harbor telomeres that are similar in size to the shortened telomeres in either single mutant, suggesting that AN3 and OLI2 genes act in the same genetic pathway for telomere maintenance. Although telomerase activity levels are wild type in all analyzed an3 and oli mutant plants, late generations of these mutants display evidence of cell death in the root apical meristem, suggesting that a common mechanism may connect telomere length control to cell proliferation and viability. Overall, our data reveal a novel set of genetic interactions governing control of telomere length and cell division in plants.

Results

Inactivation of the AN3 gene leads to shorter telomeres in A. thaliana

To evaluate the potential role of AN3 in telomere biology, we measured mean telomere length by the terminal restriction fragment (TRF) assay in two different lines of an3 mutants (Fig. 1A), and quantitated the results using the TeloTool program (Gohring et al. 2014). We first analyzed telomere length in SALK_208386c mutants (designated hereafter an3-5). These seeds were obtained from the seed stock center in the confirmed homozygous mutant state and thus were likely propagated for several generations in the absence of AN3 protein prior to our analysis. Two consecutive generations of an3-5 mutants showed shorter mean telomere length compared to the Col-0 wild type control (Fig. 1B), with an average shortening of approximately 650 bp over wild type levels (Fig. 1C). Notably, we did not detect a significant telomere size difference between these two mutant generations (Fig. 1C), indicating that telomere length in an3-5 mutants was established at a shorter, but stable, equilibrium.

Fig. 1
figure 1

Inactivation of Arabidopsis AN3 gene leads to the establishment of a new shorter telomere length set point. (A) A diagram showing the Arabidopsis AN3 gene model. Blue boxes represent exons, grey boxes represent untranslated regions, lines represent introns. Relative positions of T-DNA insertion sites and the corresponding mutant line numbers are indicated. The conserved and functionally important synovial sarcoma associated SYT N-terminal homology (SNH) domain is indicated on top. (B) TRF southern blot for Col-0 wild type and plants from two consecutive generations of an3-5 homozygous T-DNA mutants. Asterisks denote mutant plant generations that were propagated in the lab. Molecular weight DNA markers (in kb) are shown on the left. (C) Telomere length (mean TRF) distributions in ≥ 3 biological replicates of each plant genotype and generation are shown in boxplots. Data points represent mean TRF values from individual plants (biological repeats) analyzed with TeloTool. Error bars indicate minimum to maximum values. Significance codes based on unpaired t-test with Welch correlation: *P ≤ 0.05. (D) A representative TRF blot for telomere length analysis of several consecutive generations of homozygous mutant plants segregated from self-pollinated heterozygous an3-6+/− parent. (E) Telomere length (mean TRF) distributions in ≥ 3 biological replicates of each an3-6 mutant generation are shown in boxplots

Full size image

To test how many plant generations are required to reach this length set point, we analyzed a second an3 mutant T-DNA line (SALK_150407), designated hereafter an3-6. TRF analysis was performed on sibling plants segregated from self-pollination of a heterozygous an3-6+/− parent. Interestingly, telomere length in segregants of the first true an3-6 mutant generation (G1) was 3,257 bp and not significantly different from wild type siblings (Fig. 1D, E), implying that telomere loss is not immediate and may occur in later mutant generations. Indeed, mean telomere length in the second homozygous an3-6 mutant generation (G2) decreased to ~ 2.5 kb (Fig. 1D, E). Telomere length slightly fluctuated in the next mutant generations but were generally maintained in the 2.2–2.7 kb size range through at least the fifth mutant generation (G5). This newly established mean telomere length range is similar for an3-6 and an3-5 mutants, suggesting that later generations of an3 mutants maintain telomeres at a set point that is 16.6% shorter than in the wild type. We conclude that the AN3 gene is a positive regulator of Arabidopsis telomere length.

To verify that the downstream telomere effects of an3 deficiency impact different plant tissues, we measured telomere length in the shoot apical meristem (SAM) cells, where AN3 is typically not expressed (Nozaki et al. 2020). Since the TRF method for telomere length measurement requires large quantities of bulk genomic DNA taken from the entire plant (Abdulkina et al. 2023), here we utilized an alternative protocol called PETRA. This PCR-based assay is precise, requires very little genomic DNA, and relies on telomere- and chromosome arm-specific subtelomeric primers for measuring telomere length on individual chromosome arms (Heacock et al. 2004). Telomere length was analyzed in SAM of wild type and an3-6 plants on three representative chromosome arms (1L, 1R and 5R). Telomere length was shorter on all 3 chromosome arms in an3-6 mutants compared to the wild type (Supplemental Fig. 1), with mean telomere length values for individual chromosome arms being shorter on average by 535–651 bp, depending on the specific arm analyzed. These values are similar to the rate of shortening detected by the TRF assay using genomic DNA from the whole plants (Fig. 1). Overall, these findings indicate that telomere shortening is a downstream effect of AN3 inactivation and can be detected in different plant tissues.

AN3 and OLI2/NOP2A mutants have similar effects on telomere length

Both AN3 and OLI2/NOP2A were previously implicated in cell proliferation (Fujikura et al. 2009), and our data suggest that both genes are also involved in telomere length control (this study; Abdulkina et al. 2019). We thus asked if AN3 and OLI2/NOP2A function in the same or different genetic pathways for telomere length regulation. Specifically, we hypothesized that if AN3 functions in a pathway separate from OLI2/NOP2A, the addition of the an3-6 mutation would increase the rate of telomere shortening observed in oli2-4 mutant background. To test this prediction, we generated double heterozygous F1 progeny, and propagated them to F2 to obtain individual mutant genotypes and wild type controls (Supplemental Fig. 2).

We first compared telomere length in F2 siblings that were homozygous mutants for either AN3 or OLI2. Telomere length in G1 an3-6−/− plants was not significantly different from that observed in G1 oli2-4−/− mutants (which in our setup were also heterozygous for an3-6+/− mutation) (Fig. 2A, compare lanes 1,2 and 3,4; Fig. 2B). Since we could not obtain double homozygous an3-6−/−oli2-4−/− mutants in the F2 generation, we self-pollinated an3-6+/−oli2-4−/− plants to generate F3 mutants that represent the first generation of double homozygous mutants oli2-4−/− (G2) an3-6−/− (G1). Notably, these F3 double mutants displayed telomeres that were only slightly shorter relative to their oli2-4−/− (G1) an3-6+/− F2 parents (Fig. 2B, 2,010 bp vs. 2,117 bp, respectively). This finding suggests that the an3-6 mutation does not make a significant contribution to the rate of telomere shortening observed in oli2-4−/− mutants. Consistent with this observation, telomere length in the following two generations (F4 and F5) of the double mutants an3-6−/−oli2-4−/− continued to fluctuate in the ~ 1.9-2 kb range (Fig. 2A, B), which is characteristic of the late generations of oli2-4−/− mutants (Abdulkina et al. 2019). These data imply that the rate of telomere length shortening in the double oli2-4−/−an3-6−/− mutants is similar to the single oli2-4 mutant. We surmise that AN3 and OLI2/NOP2A genes, despite both being positive regulators, do not show synergistic effects on telomere length.

Fig. 2
figure 2

AN3 and OLI2 likely act in the same genetic pathway for telomere length regulation. (A) A representative TRF blot for F2-F5 generations of double mutant an3-6 oli2-4 plants. Molecular weight DNA markers (in kb) are shown on the left. (B) Telomere length (mean TRF) distribution in ≥ 3 biological replicates of each analyzed mutant genotype and generation are shown in boxplots. Data points represent mean TRF values from individual plants (biological repeats) analyzed with TeloTool. Significance codes based on unpaired t-test with Welch correlation: *P ≤ 0.05

Full size image

Mutations in AN3 or the OLIGOCELLULA cell proliferation genes do not affect telomerase activity

Given that mutations in the AN3 and OLIGOCELLULA genes (OLI2/NOP2A, OLI5/RPL5A and OLI7/RPL5B) result in a shorter telomere phenotype, we considered the possibility that these plants have reduced levels of telomerase activity. To measure telomerase activity, we performed quantitative real-time telomeric repeat amplification protocol (qTRAP) on flowers (Song et al. 2019). This assay measures enzyme activity levels using protein extracts that are rich in telomerase (Fitzgerald et al. 1999). As for wild type plants, we detected robust telomerase activity in all samples prepared from an3-6 mutants (Fig. 3). Likewise, no significant changes in telomerase activity were detected in extracts from oli5-3, oli2-4, and oli7-2 mutants (Fig. 3), indicating that inactivation of these genes does not affect telomerase activity levels in vitro. We conclude that telomere shortening in an3, oli2/nop2a, oli5/rpl5a and oli7/rpl5b mutants is not caused by biochemical changes in the telomerase enzyme activity.

Fig. 3
figure 3

Relative telomerase activity in an3 and oli mutants is similar to the wild type level in vitro. The top panel represents a histogram of the relative telomerase activity detected by real-time TRAP in flower bundles of an3-6, oligocellula mutants oli5-3, oli2-4, oli7-2, wild type Col-0 and the telomerase tert mutants (negative control). The mean of ≥ 3 biological replicates for each plant genotype is shown as a log10 of fold change compared to wild type. Error bars indicate standard deviation. Significance codes based on Students t-test: NS: P > 0.05, ***P ≤ 0.001. The bottom panel shows fluorescence change data for four individual biological replicates of the wild type, an3-6 and oli mutants and three biological replicates of the tert mutants

Full size image

Late generation AN3 and OLIGOCELLULA mutants display evidence of cell death in the root apical meristem

Mutations in major regulators of chromosome maintenance and cellular proliferation, as well as telomere length defects, can lead to genome damage and cell death in the root apical meristem (Xie and Shippen 2018; Surovtseva et al. 2009; Wang et al. 2023). Therefore, we looked for evidence of cell death in three consecutive generations of an3-6 mutants by performing propidium iodide (PI) staining of root cells. PI stains nucleic acids of dead cells but is eliminated from live cells (Amiard et al. 2010). For wild type root apical meristems (RAM), 4% of the roots we monitored contained PI-positive cells (Fig. 4A). Similarly for G2 and G3 an3-6 plants 5.5% and 7.4% of the roots we examined contained PI-positive cells (Fig. 4B, C). However, the number of PI-positive cells jumped dramatically to 25% for G4 an3-6 mutants (Fig. 4D, H). This surge in PI staining in an3-6 plants correlates with the establishment of the new shorter telomere length set point in late generations of this mutant. We next tested if the increase in PI staining is specific to AN3 mutants or can also be observed in plants with mutations in OLIGOCELLULA genes. We found that roots of all analyzed late-generation OLIGOCELLULA mutants (nop2a/oli2-4, rpl5a/oli5-3, and rpl5b/oli7-2) harbor an increased number of PI-positive RAMs compared to wild type (Fig. 4E-I). The correlation between shorter telomeres and increased meristematic cell death in late generations of an3-6, nop2a/oli2-4, rpl5a/oli5-3, and rpl5b/oli7-2 mutants suggests a common mechanism through which telomere length control is connected to cell proliferation and genome maintenance check points.

Fig. 4
figure 4

Analysis of root tips of wild type, an3 and oli mutant plants stained with propidium iodide. (A-G) Representative images of root tips of 5-day-old WT (A), an3-6 G2 (B), an3-6 G3 (C), an3-6 G4 (D), oli2-4 G4 (E), oli5-3 G3 (F) and oli7-2 G5 (G) mutant seedlings stained with PI. Arrows point to PI-positive cells in the quiescent center of root meristems in an3-6 and oli mutants. For reference, a scheme of the root tip is shown on the top (RAM). Gray cells in the root apical meristem (RAM) indicate the stem and precursor cells around the quiescent center (QC) (in white). Numbers of PI-positive roots (H) and graphical representation (I) of the percentages of PI-positive roots for wild type and mutants are shown. *P ≤ 0.05 based on Fisher exact test

Full size image

OLI5 and OLI7 genes, but not AN3 or OLI2, contribute to chromosome stability

Mutations in chromosome capping proteins or critical shortening of telomeric DNA tracts can lead to an increased incidence of end-to-end chromosome fusions and genome aberrations (Borges et al. 2022; Song et al. 2008). To check if telomere integrity is compromised in an3 and oli mutants, we performed cytogenetic analysis of mitotically dividing cells in pistils. As expected, we saw no chromosomal segregation defects in wild type plants (Fig. 5A). Although we did detect rare incidences of lagging chromosomes (Guo et al. 2020) during anaphase in an3-6 mutants, we found no evidence of anaphase bridges indicative of chromosome fusion in either an3-6 or in oli2-4 mutants (Fig. 5B and C). Notably, however, bridged chromosomes were associated with 5.8% and 7% of the anaphases in oli7-2 and oli5-2 mutants, respectively (Fig. 5D, E and F), consistent with a role for these genes in maintaining chromosome stability.

Fig. 5
figure 5

OLI5 and OLI7 genes contribute to chromosome maintenance. (A-E) Representative images of mitotic spreads from WT (A), an3-6 G3 (B), oli2-4 G3 (C), oli7-2 G4 (D), oli5-2 G3 (E) pistils. Cells were stained with DAPI and observed using a fluorescent microscope with 63X magnification. Red arrows indicate chromosome lagging, yellow arrows indicate anaphase bridges. (F) Frequencies of anaphase bridges observed for mitotic cells from each genotype are indicated.

Full size image

To evaluate if the chromosome segregation defects observed in OLIGOCELLULA mutants correspond to telomere-to-telomere fusions, we performed telomere fusion PCR (TF-PCR), an assay that detects covalently joined telomeres using combinations of primers directed at unique subtelomeric sequences from different chromosome ends (Heacock et al. 2004). As positive and negative controls, we utilized stn1 mutants and Col-0 plants, respectively (Song et al. 2008). Abundant TF-PCR products were observed in stn1 mutants, but no or few TF-PCR products were detected with oli2-2/nop2a-2 (Fig. 6) or with an3-6, oli7-2 and oli5-2 DNA (Supplemental Fig. 3). These results suggest that the chromatin bridges observed in oli5 and oli7 mutants may not result from telomere fusions and instead may represent aberrations involving other parts of chromosomes. This conclusion is consistent with previous observations for these mutants (Abdulkina et al. 2019) indicating that telomere length in oli5 and oli7 plants exceeds the critical minimal size threshold of 1.0 kb required to avert telomere-to-telomere fusions (Heacock et al. 2004). Taken together, our data indicate that AN3 and OLIGOCELLULA genes are important for the establishment of telomere length set point and for some aspects of general chromosome stability in Arabidopsis, but are largely dispensable for chromosome end protection.

Fig. 6
figure 6

PCR amplification of telomeric end-to-end chromosome fusions using chromosome-specific sub-telomeric primers. Amplification only occurs when two telomeric regions from different chromosome arms are covalently joined end-to-end. Southern analysis of fusion PCR products is then performed using a telomeric probe. Panels show Fusion PCR results for different subtelomeric primer combinations, using wild type (WT) template DNA (negative control), DNA from stn1-1 (positive control) and nop2a-2/oli2-2 mutants. Results for a single plant are shown in each lane. The subtelomeric primers used for fusion PCR are indicated, i.e., 3 L (left arm of chromosome 3) + 4R (right arm of chromosome 4). Molecular weight DNA markers (in kb) are shown.

Full size image

Discussion

Much attention has recently been given to the non-canonical functions of telomerase and telomere-associated factors in DNA repair, transcription, and cell cycle control (Barbero Barcenilla and Shippen 2019; Hong et al. 2016; Valeeva et al. 2023). Conversely, multiple lines of evidence from different eukaryotic systems indicate that many developmental and cellular pathways beyond the canonical telomere biology genes also have pleiotropic impacts on telomere length regulation. Here we show that the transcriptional co-activator ANGUSTIFOLIA3/ GRF-INTERACTING FACTOR1 (AN3/GIF1), besides its major roles in chromatin remodeling and cell division control, functions as a positive regulator of Arabidopsis telomere length. Telomeres in two an3 mutant lines shorten to a similar level as in plants deficient for the ribosome biogenesis OLIGOCELLULA genes, and ultimately establish a new shorter telomere length setpoint. Given the previously described genetic interactions between components of the OLIGOCELLULA-dependent rRNA maturation/ribosome assembly pathway and the AN3-dependent cell proliferation mechanisms (Fujikura et al. 2009), our data imply that telomere length homeostasis is controlled by an even larger scope of pleiotropic genetic networks than previously envisioned.

Considering the well-established roles of OLI2/NOP2A, OLI5/RPL5A and OLI7/RPL5B proteins in rRNA binding and biogenesis (Bourgeois et al. 2015; Fujikura et al. 2009), we asked if telomere shortening observed in these mutants could be due to decreased activity of the telomerase. However, the levels of telomerase activity in vitro remained unaltered in an3 and all tested oli mutant plants, suggesting that the OLIGOCELLULA or AN3 pathways are not involved in the regulation of telomerase activity or stability in vivo. This observation is intriguing and suggests that other mechanisms could be responsible for telomere shortening observed in these mutants. Several possibilities can be envisioned. First, the telomerase holoenzyme can be regulated in vivo through its recruitment to the DNA substrate or through changes in its subcellular localization, including cell cycle-related trafficking (Robinson and Schiemann 2022). Given the previously proposed role of AN3/GIF1 and other GRF-interacting factors in cell cycle regulation (Lee et al. 2009), the role of these proteins in regulating telomerase access to the telomere at different phases of the cell cycle is plausible. Other potential explanations for telomere shortening include deficiencies in telomeric DNA binding proteins that regulate telomerase access. Unfortunately, we still do not have a good understanding of the full complement of such factors in plants (Fulcher and Riha 2015; Procházková Schrumpfová et al. 2016). Alternatively, defects in general protein synthesis that are characteristic of these ribosome biogenesis mutants may also impact telomere length indirectly (Garus and Autexier 2021).

One other intriguing scenario is the possible involvement of OLI2/NOP2A in posttranscriptional modification (methylation) of some RNA molecules that function in both the telomerase maturation and ribosome biogenesis pathways (Valeeva et al. 2023). In humans and yeast, the presumed NOP2 methylation target is the C5 position of cytosine 4447 in 28 S rRNA (Bourgeois et al. 2015). Although Arabidopsis NOP2A does not appear to methylate rRNA in vivo (Burgess et al. 2015), it is possible that NOP2A methylates another RNA target. At least two telomere-related long noncoding RNAs could potentially serve as NOP2A methylation targets instead of rRNA: the telomerase templating RNA subunit (Fajkus et al. 2019; Song et al. 2019), and ARRET, a C-rich telomere repeat containing RNA (Vrbsky et al. 2010). The function of ARRET is currently unknown, but its complementary strand, TERRA, has been implicated in telomere length regulation (Wang et al. 2015).

Interestingly, homozygous mutants of many telomere biology genes remain viable and fertile over several plant generations until continued shortening of telomeres reaches the minimum functional length of 400 bp − 1 kb (Watson et al. 2021; Heacock et al. 2004), ultimately triggering plant sterility and cessation of plant vegetative growth. For instance, for some mutants (ctc1, ddm1) it takes 3–6 generations to reach the terminal stage (Surovtseva et al. 2009; Xie and Shippen 2018), while for the telomerase mutants (tert) it can take up to ten plant generations to reach this threshold (Riha et al. 2001). In the an3 mutants, however, telomere length declines in the first several generations to approximately 2.3 kb but is ultimately stabilized at this shortened set point that is well above the minimum threshold, possibly explaining the lack of catastrophic genomic perturbations and continued fertility of both analyzed an3 mutants.

The degree of telomere length decline observed in an3 and oli mutants is remarkably similar to that of Arabidopsis Replication Protein A (RPA) mutants (Aklilu et al. 2020). The A. thaliana trimeric RPA protein complexes have multiple roles in DNA replication and metabolism, including functions in telomere length control, but not in regulation of telomerase activity or in chromosome end protection (Audry et al. 2015). Interestingly, telomere length in several Arabidopsis rpa single and double mutants also initially declines but eventually stabilizes at a new shorter length equilibrium which is stably maintained through multiple subsequent mutant generations (Aklilu et al. 2020). The similarities in telomere phenotypes between rpa, an3 and oli mutants suggest that cell cycle and cell division control mechanisms could potentially operate in close coordination with DNA replication and recombination machinery to sustain proper telomeric DNA length.

Although AN3 gene has a discrete pattern of expression which is particularly well-characterized in leaves (Horiguchi et al. 2005), it is remarkable that telomere length changes are observed in a variety of other adult cell types and tissues of the two analyzed an3 mutants. Indeed, uniformly shorter telomeres are detected when genomic DNA is extracted from both the entire adult plant (Fig. 1) and from tissues containing the shoot apical meristem which does not express AN3 (Supplemental Fig. 1). It should be noted that such tissue non-specific telomere phenotypes are actually characteristic of most mutants of telomere biology genes, regardless of their expression patterns. While some telomere biology genes are known to be expressed in all analyzed tissues at similar levels (KU70, POT1A, STN1, etc.), others (i.e., TERT, TAC1) have a more variable pattern with substantial expression differences across plant tissues (Supplemental Table 1). Nevertheless, inactivation of genes with both types of expression profiles leads to telomere length perturbations in the entire plant body (Riha et al. 2001, 2002; Surovtseva et al. 2007; Song et al. 2008; Ren et al. 2004).

What could be the reason for the more general effects of AN3 inactivation on telomere length observed in other tissues beyond leaves? First, earlier gene expression studies and the more recent omics analyses indicate that besides basal meristematic regions of leaf primordia, AN3 is also strongly expressed in many other adult plant tissues, such as siliques, roots, upper stems, shoot tips, flower buds, and mature flowers (Horiguchi et al. 2005; Kanei et al. 2012; Lee et al. 2014; Jun et al. 2019; Mergner et al. 2020). Furthermore, even in leaves AN3 protein shows remarkable movement across layers to AN3-negative cells (Kawade et al. 2013). Finally, previous analyses of A. thaliana and other monocarpic plants indicate that telomeres do not undergo detectable replicative shortening in differentiated adult tissues, supporting the conclusion that telomere length set point for each individual plant is established very early in development (Watson and Riha 2011), most likely in meiosis through an ATM-dependent mechanism (Vespa et al. 2007). Interestingly, AN3 participates in the development of the male and female reproductive structures and gametes (Lee et al. 2014), potentially influencing telomere length even before zygote formation. Furthermore, AN3 was also shown to be expressed at multiple stages very early in embryo development where it regulates both cell division and elongation via an AN3-MINI3 gene cascade (Meng et al. 2016). Collectively, these data support the notion that telomere length is established very early in development (when AN3 is expressed) and does not change much during later stages of plant growth, explaining the effects of an3 inactivation on telomere length across various Arabidopsis tissues.

While the precise molecular mechanism for telomere shortening in an3 plants is currently unknown, our analysis of the an3-6 oli2-4 double mutant argues that AN3 and OLIGOCELLULA genes genetically interact and likely participate in the same pathway for telomere length control. The data are also consistent with previous findings from genome-wide tandem chromatin affinity purification assays placing OLI2 gene expression under AN3 control (Vercruyssen et al. 2014). AN3 and OLIGOCELLULA genes are known to functionally interact in cell size control, and combining mutations in any OLIGOCELLULA gene with AN3 mutation leads to synergistic drop in leaf cell number and to markedly enhanced compensation in cell size (Fujikura et al. 2009). Curiously, double an3-6 oli2-4 mutants do not display shorter telomeres than in either single mutant, suggesting that synergistic effects are specific to the function of these genes in cell proliferation, but not in telomere length control. These data further imply that the AN3 and OLIGOCELLULA genes interact, but their genetic pathways are at least partially separate.

We also discovered that cells in the root apical meristems of late generation an3 and all tested oli mutants are more susceptible to stem cell damage than wild type cells. This finding is consistent with recent observations concerning the influence of telomere biology genes on stem cell biology. Stem cell renewal in the Arabidopsis RAM niche is regulated by several groups of genes, including SCARECROW and PLETHORA family members (Xiong et al. 2020). Expression of some of these genes is misregulated in an3 mutants (Ercoli et al. 2018). Interestingly, SCARECROW also regulates expression of Arabidopsis telomerase as well as STN1 and CTC1, the core components of the CTC1/STN1/TEN1 (CST) telomere replication complex. Loss of any of these factors causes root developmental abnormalities and stem cell defects (Wang et al. 2023; Surovtseva et al. 2009; Leehy et al. 2013; Song et al. 2008). Furthermore, an3 mutation decreases cell number and size in the shoot apical meristem (SAM) (Lee et al. 2009; Kinoshita et al. 2020). Finally, telomere length correlates with meristem activity and stem cell development in Arabidopsis (Gonzalez-Garcıa et al. 2015). Overall, these data point to a conserved functional role of genes regulating telomere length and protecting telomere integrity in stem cell maintenance (Hashimura and Ueguchi 2011; Wang et al. 2023).

Progressive telomere shortening can ultimately lead to genome instability and profound developmental defects. If A. thaliana telomeres fall below 1 kb in length, they are recruited to end-to-end chromosome fusions to form dicentric chromosomes which in mitotic cells can be cytogenetically visualized as anaphase bridges (Riha et al. 2001; Heacock et al. 2004). Although telomere length in oli and an3 mutants remained substantially above the length threshold (1.7–2.2 kb), we observed multiple instances of chromatin bridges in late generation oli5 (7%) and oli7 (5.8%) mutants, as well as rare cases of lagging chromosomes in an3 and oli2 mutants. Notably, TF-PCR indicated that telomeres were not engaged in covalent end-joining events. We speculate that these chromosomal segregation defects do not reflect telomere dysfunction, but rather impairment of ribosome biogenesis. Misregulation of ribosome biogenesis processes is known to have an impact on genome stability and aging processes (Kobayashi 2014; Lindström et al. 2018), as well as on cell growth and proliferation (Orsolic et al. 2015; Funk et al. 2016). The incidence of anaphase bridges in oli5 and oli7 mutants is similar to sixth generation (G6) telomerase-deficient tert mutants, in which up to 6% of all anaphases are defective (Riha et al. 2001). However, while telomeres continue to shorten in the absence of telomerase causing worsening telomere dysfunction in subsequent generations (37% anaphase bridges in G7 tert and up to 50% in G8), telomere length in oli mutants stabilizes at a new shorter setpoint, preventing further decline in telomere protection.

OLI2/NOP2, RPL5 and AN3 proteins have conserved roles in ribosome biogenesis, transcription, and chromatin remodeling throughout eukaryotic evolution (Bourgeois et al. 2015; Mathieu et al. 2003; Vercruyssen et al. 2014). Besides these well-established functions, human RPL5 and SYT (the AN3 homolog) have additional non-canonical roles in tumor suppression (Fancello et al. 2017; Ma et al. 2022; Perani et al. 2003). Similarly, our data uncovered an important and unanticipated role for the AN3 gene as a positive regulator of Arabidopsis telomere length, as well as additional roles of OLIGOCELLULA genes in plant chromosome maintenance. The high degree of evolutionary conservation in telomere maintenance mechanisms (Shakirov et al. 2022) suggests that these proteins may also contribute to chromosome stability and telomere maintenance in a broad range of eukaryotes.

Materials and methods

Plant material, growth conditions, and PCR genotyping

Seeds for Arabidopsis lines an3-5 (SALK_208386), an3-6 (SALK_150407), stn1-1 (SALK_023504) (Song et al. 2008), oli2-4/nop2a-4 (SAIL_1279_H03) (Abdulkina et al. 2019), oli2-2/nop2a-2 (SALK_129648) (Fujikura et al. 2009), oli5-2/rpl5a-2/ae6-2 (Fujikura et al. 2009), oli5-3 (SALK_023075), oli7-2/rpl5b-2 (Yao et al. 2008), and wild type Columbia ecotype (Col-0; CS6673) were obtained from the ABRC stock collection. Seeds were vernalized for three days in 4ºC, sterilized in 50% bleach with 0.5% Triton X-100, and plated on MS medium containing 50% Murashige and Skoog Basal Medium with 0.5% agar and 1% sucrose. Seedlings with two true leaves were transplanted into the soil (3:1 ratio of Pro-Mix Bio-fungicide and Profile Field and Fairway Calcined Clay) and grown at 22 °C with 16 h light/8 h dark photo period. Genotyping was performed using leaf DNA as previously described (Chastukhina et al. 2016). Primers used for genotyping and for other assays are listed in Supplementary Table 2.

Telomere length and telomerase activity analyses

Telomere length analysis was carried out by non-radioactive or radioactive telomere restriction fragment (TRF) methods (Nigmatullina et al. 2016; Fitzgerald et al. 1999). Genomic DNA from individual whole plants was extracted using the CTAB method (Doyle and Doyle 1990). DNA was digested with Tru1I (ThermoFisher Scientific) restriction enzyme, resolved in 0.8% agarose gel, transferred to nylon membrane (Hybond N+) and hybridized with 5’ end labeled [T3AG3]4 probe. Telomeric signals were scanned using either the Pharos FX Plus Molecular Imager (Bio-Rad Laboratories) or the Molecular Imager Chemidocs XRS + with Image Lab™ Software (Bio-Rad). Data were visualized by Quantity One v.4.6.5 software (Bio-Rad), and mean telomere length (mean TRF) was calculated using the TeloTool program (Gohring et al. 2014).

To measure telomere length in the SAM, we employed primer extension telomere repeat amplification (PETRA) (Heacock et al. 2004). Plants of each genotype were grown horizontally on MS medium plates for 7 days, a minimum of 3–4 individual seedlings were collected, and shoot tips containing the SAM were dissected and separated from other tissues with a razor blade as described previously (González-García et al., 2015; Chou et al. 2016). Genomic DNA from the SAM region was isolated using the CTAB method (Doyle and Doyle 1990). Telomere length was measured on three chromosome arms (1 L, 1R, and 5R) (Heacock et al. 2004). Following the Southern blotting procedure, telomere signals were scanned using a Chemidocs XRS + molecular transilluminator (Bio-Rad, USA) with Image Lab™ software (Bio-Rad, USA), analyzed using Quantity One v.4.6.5 software (Bio-Rad, USA), and the mean telomere length (mean TRF) values for each analyzed chromosome arm were calculated using the TeloTool program (Gohring et al. 2014).

Relative telomerase activity was measured using quantitative telomeric repeat amplification protocol (qTRAP) as previously described (Song et al. 2019). 3–5 inflorescent bundles from individual plants were ground in liquid nitrogen and resuspend in buffer W+ (1 M Tris–Acetate pH 7.5, 1 M MgCl2, 2 M KGlutamate, 0.5 M EGTA, 1.5% PVP, 10% glycerol, 1 µM DTT, 0.6 nM RNAse inhibitor (Sintol E-055), 1 µM Protease Inhibitor Cocktail (Sigma). Protein concentration was measured by DC™ Protein Assay (Bio-Rad). Primer extension step was performed with primer qTRAP-F for 45 min in the dark at room temperature, followed by the qPCR step using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) with primer qTRAP-R. The average level of telomerase activity in the wild type plants (Col-0) was set as 1. At least three biological replicates (3–5 flower bundles from one plant per replicate) with three technical replicates were performed for each analyzed plant line (Supplemental Data 1).

Cytogenetic and telomere fusion PCR assays

For mitotic analysis, spreads were prepared from pistils and stained with DAPI (40,60-diamidino-2-phenylindole), as previously described (Riha et al. 2001). Spreads were imaged at 63X in oil using the fluorescent microscope LSM 780 (Zeiss, Germany). Telomere fusion analysis was performed by PCR using 2 µg of genomic DNA and specific subtelomeric primers as previously described (Heacock et al. 2004). For the extension step, ExTaq polymerase (Takara) was used, and PCR products were resolved in 1% agarose gel with further transfer to the membrane and hybridization with the telomeric probes as described for the TRF analysis. The presence of fused chromosomes was analyzed using subtelomeric primers specific to chromosome arms 1 L, 1R, 2R, 3 L, 3R, 4R, 5 L and 5R, as described before (Heacock et al. 2004).

Root cell death assay

After germination, seedlings were grown vertically on MS media plates for 5 days, and intact roots were collected and transferred to Propidium Iodide solution (10 mg/ml in MilliQ water) for 30 s and rinsed with MilliQ water as described (Bose et al. 2020). Roots were then transferred to slides in a droplet of H20, sealed with a cover slip and imaged at 40X with dsRED filter using a fluorescent microscope LSM 780 (Zeiss, Germany). Only roots with at least a single PI-stained cell in the quiescent center of the root stem cell niche were counted as evidence of cell damage in RAM.

Statistical analysis

Statistical analysis for telomere length was carried out with GraphPad Prism 8 software using unpaired t-test with Welch correlation. The difference in the values between the wild type (control) and the mutant (experiment) was considered significant for p ≤ 0.05. For the qTRAP experiment, relative telomerase activity was normalized to the average Cq value (meanCq) of a positive control (wild type) using the log2(Cq-meanCq) formula. For analysis of cell death in RAM, statistical significance was calculated using Fisher exact test with the GraphPad Prism 8 software.