The Prominent Role of Translation Factors, Heat shock Protein 90-2 and an ABC Transporter Associated to the Resistance of a Maize Line against Sugarcane Mosaic Virus

: The mosaic disease in maize is caused by Sugarcane mosaic virus (SCMV), a member of the Potyviridae family. The best strategy to cope with viral infections is the use of disease-resistant maize lines. To better understand the resistance response to SCMV, we analyzed differentially expressed genes among a resistant line (CI-RL1), a susceptible line (B73), and the F1 progeny from a cross between both lines using RNA-Seq data. We also analyzed transcript expression pattern clustering to allocate previously reported resistance candidate genes. GO enrichment analysis of biological processes highlighted a strong regulation in ROS detoxification in both the susceptible and resistant lines. The enrichment of cellular components led to the identification of an integral component of the plasma membrane in the RL line. Transcript expression patterns provide evidence of the importance of host translation in virus response, showing the diverse and complex behavior of eIF4E homologs and the presence of eleven eEF1α factors in maize. In addition, we identified two genes putatively implied in long-distance movement: ZmPiezo and ZmPVIP1. Finally, we propose an ABC transporter to be associated with viral resistance.


Introduction
Maize is one of the most important food and staple crops around the world. However, like other important crops, it is affected by a variety of pathogens. The mosaic disease in maize is caused by Sugarcane mosaic virus (SCMV), a member of the Potyviridae family. Disease symptoms include leaf mosaics, chlorosis, and stunting, depending on the infected maize variety. SCMV has been shown to establish synergistic infections with Machlomovirus (Tombusviridae family), causing Maize lethal necrosis (MLN) [1], with devastating effects (with losses of up to 90%) in maize cultivars in China and Africa [2,3]. The best strategy to cope with such viral infections is the use of resistant maize lines. However, the mechanism underlying maize resistance remains poorly understood.
The defense against plant pathogens relies on the innate immunity of each plant cell and on the systemic signals produced at the infected sites [4][5][6][7]. Plant innate immunity is activated after the recognition of microbial-or pathogen-associated molecular patterns (PAMPs and MAMPs) by transmembrane recognition receptors (PRRs) [6,7]. PAMPs-triggered immunity (PTI) is a general and non-specific defensive response, characterized by the activation of early responses such as production of reactive oxygen species (ROS), signaling cascade inducers (i.e., mitogen-activated protein kinases; MAPKs), and the accumulation of callose [6,8]. Pathogens can synthesize molecules that interfere with the activation of the PTI, which are called effectors [9]. If plants possess resistant (R) genes, proteins with nucleotide-binding sites (NBS), and leucine-rich repeats (LRR) domains [10], they can specifically recognize the effector, resulting in Effector-triggered immunity (ETI) [7]. ETI leads to the activation of ROS, MAPKs, and phytohormone signaling routes [11][12][13][14]. In addition, ETI also activates the production of defense proteins associated with cell death in a process called the hypersensitive response (HR) [15,16].
Another defense mechanism against viral infection is RNA silencing [17,18]. The mechanism is triggered by the detection of double-stranded RNA (dsRNA). The dsRNA is cleaved by Dicer-like proteins (DCL) into small interfering 21-25 ribonucleotides (sRNAs). One of the dsRNA strands is loaded in the RNA-induced silencing complex (RISC) by the Argonaute (AGO) proteins. The RISC uses the sRNA as a guide to direct RNA degradation [17].
Another set of genes playing a role in resistance are the susceptibility factors. These are recessive genes playing a role in the virus infection cycle [19]. When one of these factors have mutations, the host-pathogen interaction is affected, with a concomitant resistance phenotype in the host. Most of the resistant genes to potyvirus identified in important crop plants correspond to alleles of the eukaryotic initiation factor 4E (eIF4E) [20].
Susceptibility factors, other than eIF4E, include those associated with long-distance virus movement. One of them is PVIP1 (Potyvirus VPg interacting protein) [21]. Mutants in Arabidopsis thaliana (L.) Heynh. that do not allow such interactions presented restriction of the long-distance movement of the virus, thus granting resistance to the viral infection.
The resistance to SCMV in maize has been assigned to the presence of the Scmv1 and Scmv2 loci in chromosomes 3 and 6, respectively. Both loci have been further explored, and the candidate genes were isolated. Scmv1 is an atypical thioredoxin h (Trx h) [22], while Scmv2 is an Auxin-binding protein (ABP). Thioredoxins have a WC(G/P)PC motif, with two active-site cysteines (C) that participate in disulfide bridges reduction [22][23][24]. An atypical ZmTrx h has those two C substituted with asparagine (N) and serine (S), yielding a WNQPS motif. ZmTrx h possesses 67 homologs in the maize genome, of which 32 are non-redundant without the WNPQS motif. It has been proposed that the Trx hinduced resistance was due to a higher gene expression, as it was also detected in susceptible plants. A mutation in the promoter region of ZmTrx h is responsible for the increase in transcript accumulation, but the mechanism related to resistance remains unclear [22].
To better understand the resistance response to SCMV, we analyzed differentially expressed genes between a resistant line (CI-RL1), a susceptible line (B73) and the F1 progeny derived from both lines. We carried out two main approaches with transcriptomic analyses: the first involved identification of the DEGs in the functional enrichment network of cellular components (CC), while the second involved transcript expression pattern clustering to allocate previously reported candidate genes. We observed, through GO enrichment analysis of biological processes (BP), a strong enrichment of ROS detoxification related genes. Furthermore, the enrichment of CC led to the identification of an ABC transporter. Transcript expression patterns provided evidence for the diverse (and, hence, complex) behavior of the eIF4E homologs in maize-particularly for the eEF1α factors. Expression pattern analysis also allowed for the identification of two genes implied in long-distance movement: ZmPiezo and ZmPVIP1.

Symptom Development in Susceptible Line, Resistant Line, and F1 Progeny
Six plants of each line (B73, CI-RL1, and F1 cross progeny) were grown in a greenhouse until they reached the third-leaf developmental stage. Then, all of their leaves were mechanically inoculated with SCMV-Ver1 isolate (EU091075.1). The plants were observed daily for the presence of mosaic symptoms. The susceptible B73 plants showed early mosaic symptoms in the ligule zone, as soon as 7 dpi. In contrast, CI-RL1 (RL) and F1 progeny (F1) plants were symptom-delayed, compared to B73 plants. RL 1 plants were almost identical to the uninfected controls, without any symptoms or developmental alterations. The F1 plants showed only a slight mosaic near the central vein and were like healthy plants ( Figure 1A) with a slight mosaic in the youngest leaves. Chlorosis was observed in B73 plants, with a strong mosaic in the first-and second-youngest leaves, while a slight mosaic was observed in the basal region for the F1 leaves ( Figure 1B). The F1 plants appeared to be taller than the other two lines, but this could be attributed to the heterosis effect.

RNA-Seq of Virus-Infected Maize Plants
To analyze the global host response to SCMV infection, we obtained the transcriptomes of B73, F1, and RL maize lines at 17 dpi, with their respective mock-inoculated control set and in duplicates. Therefore, a total of 12 libraries were pair-end sequenced (PE 2 x 100), each one producing at least 20 million reads.

Gene Expression
Profiling is More Similar Between Susceptible B73 and Symptom-delayed F1, compared to RL l After the differential expression analysis, genes with p-value ≤ 0.01 were considered as differentially expressed genes (DEGs). We found a total of 1959 DEGs in all maize lines and, of those, 924 corresponded to B73, 633 to F1, and 583 to RL. For B73, we found 496 upregulated and 459 down-regulated DEGs; 428 were up-and 206 down-regulated for F1; and 289 were up-and 304 down-regulated in the resistant CI-RL1 line. Only eight genes were differentially expressed in all three lines, corresponding to the 0.4% of the total DEGs ( Figure 2A). The susceptible B73 line and the symptom-delayed F1 line were observed to be the most similar, sharing 5.7% and 3.2% of the up-and down-regulated genes (65 and 29 DEGs), respectively ( Figures 2B and C). The resistant RL line shared only 1.1% and 2.3% of the total DEGs with F1 and B73, respectively. This feature-that is, fewer genes differentially expressed in the resistant line-is consistent with other transcriptomes when comparing resistant and susceptible lines. There are differences in the total (Figure 2A) and the addition of the up-down-regulated DEGs ( Figures 2B and 2C) for each line. This is due to the different number of transcripts considered for each gene, behaving differently in each line, i.e., transcripts of the same gene being up-or down-regulated. Thus, the same gene will appear in both 2B and 2C diagrams and, when added, the total number will be higher than in the 2A diagram.

Membrane-Associated Cellular Components are Up-and Down-Regulated After SCMV Infection
To better understand the transcriptomic response of our three different maize lines to SCMV, we performed a functional enrichment analysis. A salient feature of the cellular component (CC) analysis enrichment was up-regulation of genes for the integral component of the plasma membrane in the RL line ( Figure 3A). For B73 + F1, nine cellular processes were up-regulated: nuclear outer membrane-endoplasmic reticulum membrane network, cytosol, endoplasmic reticulum membrane, cytosolic ribosome, ribosomal subunit, proteasome complex, intracellular membrane-bounded organelle, non-membranebound organelle, and ribonucleoprotein complex. Five CCs were up regulated for B73+F1+RL: integral component of membrane, intrinsic component of plasma membrane, intrinsic component of membrane, cytoplasm, and intercellular organelle. Three CCs were up-regulated for B73+RL: microtubule, membrane-bound organelle, and nucleus. Most of the CCs were related to the network. On the other hand, nine down-regulated CC genes were found to be enriched only for B73 ( Figure 3B), where four of them were connected and four did not have connections. Three down-regulated CCs were enriched for B73+F1+RL: integral components of membrane, intrinsic components of membrane, and membrane. They were connected only with intracellular anatomical structure for B73+RL ( Figure 3B). Nodes represent an enriched CC ontology term and edges represent the relationships between these ontologies. CC shared terms correspond to: B73 alone (dark green); between B73 and F1 (light green); between B73 and RL (red); shared between all lines (purple); and those corresponding to RL alone (orange). Four biological processes (BPs) in the enrichment analysis were most up-regulated for B73 and F1 lines: translation, rRNA metabolic process, ribosome biogenesis, and ribonucleoprotein complex sub-unit organization. Four processes were also up-regulated for only B73: membrane fusion, hydrogen peroxide catabolic processes, chaperone-mediated protein folding, and cellular oxidant detoxification ( Figure 4A). Fewer up-regulated genes in the RL line were related to protein metabolism and protein activation (regulation of protein metabolic process, regulation of phosphorous metabolic process, regulation of molecular function, protein-containing complex sub-unit organization, phosphorylation, and regulation of protein metabolic process).
Down-regulation in BPs was observed in the resistant line (RL) for translation elongation, and that in translation initiation was observed for B73 ( Figure 4B). Regulation of hydrolase activity seemed to be equally down-regulated for both B73 and RL lines. The most striking down-regulation of BPs was in chaperone-mediated protein folding for B73 and cellular oxidant detoxification for RL.

Resistance Candidate Genes: Mostly present in F1 and Down-regulated in B73.
We looked for previously reported resistance candidate genes in the literature within our set of enriched BPs. Then, we selected some genes according to their up-or downregulated expression levels. In Table 1, we present 19 up-and 24 down-regulated selected candidate DEGs. The first set corresponds to eight down-regulated BPs: translation elongation, cellular detoxification, biological regulation chloroplast, response to external stimuli, regulation of gene expression, carbohydrate metabolic process, and digoxygenase activity. The second set (up-regulated) are within four enriched BP terms: translation, RNA metabolic process, gene silencing by miRNA, and response to stress. Within the set of 24 down-regulated genes, 16 are for the B73 line, with Log2 FC values ranging from -1 to -6 (for Argonaute 1b and Heat shock 70 kDa protein 17, respectively), whereas only four were down-regulated for RL (Translation elongation factor EF1A, Elongation factor 1α, Thioredoxin superfamily protein, and jasmonate-regulated gene 21). For the up-regulated set of genes, 12 were associated to the F1 line, 6 to B73, and 4 to RL. Translation was the mostly represented biological process, with five genes: elongation factor 2 and its sub-unit β, elongation factor 1α, translation initiation factor 4G, and translation initiation factor 3 sub-unit M. GO software sorts the genes in hierarchical groups. Therefore, a BP can encompass another BP explaining why the same gene can be found in just one or shared between two or more GO terms. This is why Thioredoxin superfamily protein is in two GO hierarchical terms, the GO:0065007-biological regulation and the GO:1990748-cellular detoxification. Meanwhile, another thioredoxin annotated as Thioredoxin-like 4/Thioredoxin-like protein CXXS1 is only found in GO:0065007-biological regulation (Table 1) reflecting annotation and enrichment of GO differences. We have no evidence of the presence of ZmTrx h within this group. Both argonaut1 and argonaut1b, were found in the GO:0035195-gene silencing but only argonaute1b found in GO:0065007-biological regulation. None of them were associated with resistance to SCMV.

Most of the Candidate Genes Have Higher Basal Expresion in RL
Analysis of the DEGs highlighted their possible role in SCMV resistance; for instance, eEFs, dicer-like, and argonaute were detected. However, this approach may have excluded recessive resistance genes or susceptibility factors with unchanged expression levels. In order to detect these types of genes, we grouped transcripts into patterns. Five expression pattern clusters were generated, showing basal expression levels of genes for each of the three evaluated lines. The pattern clusters were then grouped according to the basal level of expression for the RL line, thus resulting in groups 1 and 2 ( Figure 5A). Group 1 had a high basal expression level for RL and included clusters C0, C1, and C2. Group 2 had a low level of basal expression for RL and contained clusters C3 and C4. Considering previous attempts to understand the resistance of maize to SCMV, a list of 57 genes involved in potyvirus resistance were extracted from bibliographic data [25,26]. We then searched for the list of 57 candidate genes in all the clusters and found 43. Of these, there were 13 in group 1: two of them were associated with long-distance movement (PVIP, and ZmPiezo), three with cell-to-cell movement (BG3, Pcap1, and Myosin), and seven with the initiation or enhancement of viral replication (EXA1, LSM1, PpDDXL, IRE1A, RIM1, Chl-PGK, and RISP). Additionally, one of the detected genes (SAMS1) does not partipate in the virus infection cycle. Only three genes were found exclusively in group 2 (IRE1B; PDLP in C4; and PBAP8 in C3); see Figure 5B. The remaining 27 candidates were found in cluster combinations, as they had high and low levels of basal expression for RL ( Figure 5B). This grouping helped us to understand the significant/relevant differences in transcript expression between the resistant line (RL) and the susceptible line (B73). eIF4E was also observed in the expression pattern clusters, due to its importance in the potyvirus infectious cycle and its role as a susceptibility factor. Two candidate genes were identified in the maize genome: the first was annotated as eif6, and the second corresponded to the gene model GRMZM2G113096 (with no annotation and termed ZmeIF4E). The expression of eif6 had a low basal level in the F1 and RL lines, while eif7 and ZmeIF4E had a low basal level in the RL line.

Differences in Translation and Elongation Factor Coding Transcripts as the Possible Source of RL Resistance
The interaction of the potyvirus VPg with the translation initiation factor eIF4E and/or its isoforms plays a central role in the virus infectious cycle [27]. A search in the maizeGDB database revealed that maize possesses six genes annotated as eukaryotic initiation factors ( Table 2). In a prior search for homologs of the eIF4E from Arabidopsis in maize, the best candidate identified in the maize genome corresponded to the gene model GRMZM2G002616, annotated as eif6. As this factor was not found to be differentially expressed, its sequence was used to search within the expression pattern clusters. eif6 was found in the C4 cluster, with low levels of basal expression in both the RL line and the F1 progeny, while having high expression levels in B73. Additionally, two more eIFs were identified in cluster C3: one corresponding to the gene model GRMZM2G022019 and annotated as eif-7, and other corresponding to the gene model GRMZM2G113096 without annotation and assigned as ZmeIF4E.
In contrast, the eukaryotic elongation factors (EEFs), which also participate in protein production, presented interesting results. First, we found four differentially expressed gene models related to eEF1α: GRMZM2G001327, annotated as elfa12; GRMZM2G154218m annotated as elfa3; GRMZM2G343543, annotated as elfa10; and GRMZM5G850607, not annotated and designated as ZmeIF4E. Furthermore, the behavior of each transcript was unique, and two were up-regulated: elfa10 in F1 progeny and the B73 line, while elfa3 only in the B73 line. In contrast, elfa12 and Zm-elfa showed downregulation only in the RL line.
A further search in the MaizeGDB database for annotated eIFs and eEFs showed that maize possesses at least 6 genes annotated as eIFs and 11 genes annotated as eEFs (Supplementary Table 1). Additionally, as observed in the DEGs data and the transcript expression pattern clustering analysis, different gene models and different transcripts of the eIFs and eEFs were found, depending on the genetic background. Thus, the results of the response to SCMV could be the effect of the transcripts favored in each individual. Table 1. Up-and down-regulated candidate genes from enriched BPs. Down-(in blue) and up-regulated (in red) genes, their GO terms, maize transcript IDs, fold-change expression levels (Log2), and associated annotation are presented. *The selected candidate from the integral component of plasma membrane enriched CC corresponding to an ABC transporter.  Table 2. Transcripts coding for translation/elongation factors found in clusters C0 to C4 (see Figure 5A).  Gene  Transcript ID  Gene  Transcript ID  Gene  Transcript ID  Gene  Transcript ID  Gene  Transcript ID  elfa1  Zm00001d009868_T004  elfa1  Zm00001d009868_T009  elfa1  Zm00001d009868_T003 ZmeIF4E Zm00001d041973_T001 eif-6 Zm00001d041682_T001  elfa2  Zm00001d037875_T005  elfa2  Zm00001d037875_T008  elfa1  Zm00001d009868_T012  eif-7  Zm00001d014065_T001 elfa2 Zm00001d037875_T006  elfa3  Zm00001d037873_T011  elfa3  Zm00001d037873_T009  elfa2  Zm00001d037875_T003  elfa1  Zm00001d009868_T002 elfa3 Zm00001d037873_T001  elfa9 Zm00001d046449_T004 elfa10 Zm00001d036904_T002 elfa11 Zm00001d037905_T003 elfa12 Zm00001d037877_T005 elfa7 Zm00001d009870_T013 elfa11 Zm00001d037905_T002 elfa10 Zm00001d036904_T005 elfa12 Zm00001d037877_T008

Discussion
Maize resistance to SCMV was reported more than 30 years ago [28]. However, two genes proposed for resistance-thioredoxin and nABP, corresponding to the loci Scmv1 and Scmv2, respectively-have recently been discovered. The selected line, CI-RL1, used in this work was identified in inbred evaluations for resistance to potyviruses at CIMMYT. We have previously shown that long-distance movement of SCMV is impaired in this line [29] and, therefore explored the possibility of potyvirus movement restriction involved in resistance, along with other 43 selected candidate genes. The latter were searched for in a transcript expression pattern clustering and/or in DEGs derived from their expression using RNAseq analysis.

DEGs and GO Pinpoint Viral Replication Complexes, detoxification and chaperones
As observed in planta, the presence and severity of symptoms were more intense in B73. Correspondingly, most of the differences in DEGs related to BPs were found in this maize line and were associated with the homeostatic equilibrium maintenance response, such as membrane fusion, peroxide hydrogen catabolic process, chaperone-mediated protein folding, and cellular oxidant detoxification ( Figure 4A).
An increase in the membrane fusion BP is not surprising during potyviral infection. The formation of vesicles, such as viral replication complexes (VRCs), is an essential step in this interaction. Inside these vesicles, host cellular components are recruited and used for virus replication [30,31], along with dsRNA intermediaries, which have been suggested to subvert the RNA silencing-based defense mechanism [32][33][34]. Some potyviruses use the chloroplast and ER membranes [34,35] to produce VRCs, with concomitant chlorophyll breakdown and leaf yellowing during severe chlorosis [36][37][38], which was also consistent with the GO term enrichment analysis ( Figure 4A). In fact, this could also explain the increase in the number of genes involved in the hydrogen peroxide catabolic process and in cellular oxidant detoxification. The liberation of fragments or the content of some organelles could act as signals in the attack of a pathogen in the so-called damageassociated molecular pattern (DAMPs), generating a ROS burst response [39][40][41][42]. The over-accumulation of ROS species (also called oxidative stress) can damage and disrupt the function of cellular components [43,44]. Under normal physiological circumstances, the production and degradation of ROS species are in a steady state. Therefore, due to the constant production of ROS species in the cell, detoxification of the inner cell environment is necessary. This could explain the increased expression levels in RNA-seq data of genes involved in these processes (Figures 4A and B). The cellular oxidant detoxification genes were up-regulated in the resistant CI-RL1 line. This up-regulation could be explained as an effective and transient defense response to prevent further cellular damage. Thus, this increase appears to be more a consequence of an early response to infection, rather than a strategy used for resistance.
Chaperones associated BP were also enriched. They participate in protein folding, enabling mis-folded or aggregated proteins to correctly fold [45,46], and are also involved in targeting proteins for degradation [47]. They are also activated in response to biotic and abiotic stresses. Chaperone-mediated protein folding was up-and down-regulated in the B73 line (Fig. 4B). Potyviruses recruit the heat-shock protein 70 (Hsp70) chaperone in the formation of replication complexes [48,49]. Another pair of Hsps-Hsp40 and Hsp90appear to be involved in viral infections [50][51][52]. We detected the up regulation of the Heat shock protein 90-2, only in the RL line (Table 1). Furthermore, we corroborated this increase in its expression level by qRT-PCR ( Figure S4) suggesting its association with resistance.

Integral ABC Transporter in Resistant Maize Line
The only component enriched for the resistant RL line belonged to the integral components of the cellular membrane. One candidate was especially noteworthy: the ATPbinding cassette (ABC) transporter corresponding to the maize GRMZM2G085111 gene model (Table 1). ABC transporters move diverse structural unrelated components across the membrane and intervene in plant-pathogen interactions [53][54][55][56]. A wheat ABC transporter (lr34 gene) has been reported as a multi-disease resistance gene (MDR) against fungal pathogens [57,58]. Furthermore, maize plants expressing the Lr34 wheat gene showed resistance against two fungal pathogens [59]. The protein codified by the maize gene model GRMZM2G014282 has been identified as the best homolog to the Lr34 protein. The GCN-type ABC transporter from Lilium regale E.H.Wilson has been shown to be involved not only in the defense against fungi, but also against viruses [56]. This pathogen resistance is not due simply to the presence of the ABC transporter, as has been discussed by Sun et al. (2016) [56]: a more complicated network of genes, in which the transporter plays a pivotal role, appears to be involved. Overall, GRMZM2G085111 deserves further research.

Genes Involved in Virus Replication Found in Cluster Pattern Analyses
We found two candidates associated with virus replication (in group 2, Fig. 5): Poly-A Binding protein 8 (PABP8) and inositol-requiring protein 1B (IRE1B). The PABPs bind to the 3´-end of mRNA. Three (PABP2, PABP4, and PABP8) out of eight isoforms in Arabidopsis showed increased protein and mRNA levels during TuMV infection, besides the interaction of PABP2 with the VPg and RdRp of TuMV. Also, in pabp2pabp4 and pabp2pabp8 double mutants, a reduction in TuMV mRNA levels was observed [60]. While homologs for the three AtPABPs were searched, only PABP8 was detected in Group 2 of C3 cluster (Table in Figure 5B), possibly due to its implication in virus accumulation.
IRE1B and IRE1A homologs in Arabidopsis are involved in the splicing of bZIP60 in the response to biotic and abiotic stresses, resulting in the activation of stress-related proteins. The ire1aire1b double mutant inoculated with TuMV showed delayed symptoms and low virus accumulation [61]. Thus, both IRE1 genes are necessary in Arabidopsis to confer resistance. In maize, IRE1A was found in group 1, while IRE1B was found in group 2. As IRE1A and IRE1B appears to act redundantly, they may not participate in RL resistance.
The candidate genes EXA1, LSM1, PpDDXL, IRE1A, RIM1, Chl-PGK, and RISP (Table S2), with high levels of basal expression in the RL line (group 1), play roles in viral replication and, therefore, were not considered to be associated with resistance.

Cell-to-cell Movement-Related Genes
Viruses move from cell to cell through plasmodesmata (PD), which regulate the size exclusion limit (SEL) through callose accumulation and the participation of viral movement proteins (MPs). Potyviruses, not having a specialized MP, rely on the CI, CP, HC-Pro, and VPg proteins for cell-to-cell movement, inducing the formation and movement of VRCs, via the cytoskeleton to and through the PD, in association with CI [62]. Therefore, proteins that modify or interact with the PD can be considered important candidates in cell-to-cell movement. We found candidates involved in cell-to-cell movement in group 1 (BG3, PCap1, and Myosin) and in group 2 (PDLP) (Table in Fig. 5B).
The first of the candidates in group 1, β-1,3-glucanase (BG3), was found in the three clusters of groups 1, in Fig 5B (C0, C1, and C2). BG3 encodes an enzyme that degrades callose deposited in the plasmodesmata [63]. Callose deposition and its degradation are controlled by two enzymes: callose synthase (CalS) and BG3, respectively. It has been suggested that callose deposition restricts the spread of pathogens [64][65][66]. It is tantalizing to speculate that the oxidative stress indicated by the RNAseq data could have led to callose accumulation in the PD [67]. This accumulation requires the activation of BG3 for its degradation; however, experimental support is required to confirm such an association.
The second candidate of group 1 in Fig. 5 was the plasma membrane-associated cation-binding protein 1 (PCap1). The expression level of PCap1 in Capsicum annum L. has been shown to be related to Potato virus Y accumulation and cell-to-cell movement [68]. Furthermore, protoplasts with a pcap1 loss-of-function mutation in A. thaliana accumulated high levels of TuMV, discarding its participation in virus replication [69]. Additionally, the interaction between PCap1 and the P3N-PIPO from TuMV appears to be involved in the localization of the CI complex at PD [62,69]. As PCap1 favors a possible increase in cell-to-cell transport, its involvement in SCMV resistance appears unlikely.
The third candidate gene from group 1 was Myosin, a motor protein involved in a variety of mobility processes. As such, the silencing of Myosin XI-2 from Nicotiana benthamina Domin has been shown to inhibit the movement of the Tobacco mosaic virus (TMV), but not for Potato virus X (PVX) and Tomato bushy stunt virus (TBSV) [70]. Apparently, as plants possess different genes that code for myosin, viruses interact selectively with them. This Myosin homolog was found to have a high expression level in the RL line, which could favor SCMV accumulation and movement, thus not enhancing host resistance.
The only candidate in group 2 associated with cell-to-cell movement was a PD-located protein (PDLP) belonging to a family of type-I membrane proteins. These proteins travel along the secretory pathway to reach the plasma membrane inside the PD [71,72]. PDLPs interact with viral movement proteins (MPs), which are capable of assembling as tubules within the PD [71]. As potyviruses do not belong to the group of tubule-forming viruses, the role of PDLP in SCMV-infection remains uncertain.

Long-distance Movement-Related Genes
Regarding long-distance movement and the need to systemically infect the host, only two candidates were found in group 1 (Fig. 5): the Piezo and PVIP homologs. The ZmPiezo gene was chosen from the maize genome as an ortholog to the ESC1 gene from Arabidopsis. ESC1 codes for an ion channel, PIEZO, that responds to mechanical stimuli [73]. The esc1 mutant showed alteration in the long-distance movement of TuMV in A. thaliana [74]. The ESC1 gene, further called AtPiezo, was transcriptionally induced by viral infection [74]. The ZmPiezo gene was found to have a high level of basal expression in the RL line. Not know much about ZmPiezo in maize, so it is difficult to say whether its high level can be associated with SCMV resistance.
Considering the Potyviral VPg Interacting Protein 1 (PVIP1), the pvip1 loss-of-function mutant in A. thaliana inoculated with TuMV resulted in aberrant phenotypes and a lack of viral symptoms in planta [21]. This could be interpreted as the need for PVIP1 to ensure viral infection. Here, not only the RL line has this gene, but it was found in high basal levels ( Figure S4) with possible SNPs.

The Presence of Diverse Transcripts for eIFs and eEFs Imply a Complex Regulation of Translation in SCMV Interaction with Maize Transcripts in Resistance
Of paramount importance in this work is the role that translation seems to be playing in RL resistance involving factors in its initiation step. This was not surprising as several of these initiation factors have been well documented during the last decades. However, little is now for the role of elongation factors in viral infection. Eukaryotic initiation factors (eIF), had negative regulation in BPs for the B73 line ( Figure 4B). It appears that the expression level of eIF4E is important for the establishment and maintenance of viral infection. Plants that transiently overexpress a modified eIF4E showed an increase in resistance to Potato virus Y and a decrease in the expression level of the host gene [75]. Inhibition of cap-dependent translation, by the interaction of the host eIF4E with VPg, favors the use of Internal Ribosomal entry sites (IRES) in the 5' UTR region of some potyviruses [76]. Furthermore, different kinds of stresses appear to stimulate IRES translation [76,77] and down-regulate the production of new host proteins [78], thus influencing the eIFs. This would be the case for the stress caused by SCMV infection.
The ortholog for the eIF4E from Arabidopsis (AT4G18040) corresponds to the maize eif-6 transcript (Zm00001d041682), one of the six genes annotated as eukaryotic initiation factors in the maizeGDB database (Table S1). Transcript eif-6, within cluster C4 (Table 2 and Fig. 5A), has low basal expression levels in the resistant line (RL) and the late-symptom F1 line, implying its participation in the virus infection cycle. On the other hand, eif-7 and ZmeIF4E had low basal expression levels only in the RL plant. The best match in the Arabidopsis genome for the eif-7 genes corresponds to the isoform of the eIF4E factor. The use of the eIF4E or its isoform depends completely on the interaction of the host and the infecting virus. Both eif-7 and ZmeIF4E can be newly described alleles in maize associated with resistance, for instance, restricting viral movement in the RL line, as has benn described for TEV in A. thaliana [79]. Both B73 and RL had different basal levels of ZmeIF4E (Table 2), suggesting either differences in their coding or in the promoter region for this gene.
The eukaryotic elongation factors (eEFs) are encoded by genes playing a central role in the elongation step of translation [80]. One of them-namely, eukaryotic elongation factor 1α (eEF1α)-is a cytoplasmatic protein that delivers aminoacylated tRNAs to ribosomes during polypeptide elongation when bound to GTP [81,82], and play a role in nuclear export, proteolysis, and apoptosis [83]. In human viral infections, eEF1α has been shown to interact with the Gag1 polyprotein of the Human immunodeficiency virus type 1 (HIV-1) [84], the 3´stem-loop region of the West Nile virus (WNV) [85], and with the nucleocapsid protein of the Paramyxovirus Respiratory Syncytial Virus [86]. In plants, eEF1α has been shown to bind tRNA-like structures (TLSs) in the 3' region of Turnip yellow mosaic virus (TYMV), Tobacco mosaic virus (TMV), and Brome mosaic virus (BMV) [87,88]. The aminoacylation of the TLSs enhance virus protein translation and facilitates virus RNA encapsidation [89][90][91]. The interaction between the ER-localized P3 from the Soybean mosaic virus (SMV) and eEF1α has been shown to be essential for virulence in the susceptible host [82]. Furthermore, eEF1A from N. benthamiana has been identified as a pro-viral factor required for the Tomato spotted wilt virus disease (TSWV) [92]. Maize possesses at least eleven eEFs, from which nine are annotated in the MaizeGDB database as "elongation factor α (elfα)" (Supplementary Table 1). A variety of functional transcripts are associated to each of them; for example, elfa1 has 15 transcripts associated with this gene model. In our case, we speculate that high expression levels of elfa10 and elfa3 (Table 1, red) suggest these factors interact with SCMV in a yet unforeseen manner. In contrast, RL showed down-regulation in elfa12 and Zm-elfa (Table 1, blue). It is possible that elfa10 and elfa3 interact better with SCMV, maintaining low expression of the non-interactors elfa12 and Zm-elfa. Additionally, according to the data obtained in the expression pattern cluster profiles, different transcripts of the same gene model with high or low basal levels could be found among the clusters, suggesting an intricate regulation of these factors in maize during SCMV infection. It will be necessary to dissect the interaction of elfa10, elfa3, elfa12 and Zm-elfa with SCMV to disentangle the myriad of translation factors involved in this hostvirus interaction.
Besides the large number of possible important genes for resistance, SNPs also need to be considered for some of the candidate genes discussed here, adding more complexity than anticipated in maize-SCMV interactions. Finding them will contribute to breeding programs against devastating virus diseases in monocots.

Virus and Plant Matherials
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 31 January 2022 doi:10.20944/preprints202201.0464.v1 Sugarcane mosaic virus isolate Veracruz 1 (SCMV-VER1) was used as the viral inoculum. Biologically infected maize tissue collected from symptomatic plants in Veracruz fields was stored in an ultra-freezer at -70 °C. The inoculum was reactivated by inoculated B73 susceptible plants with a mix of approximately 500 mg of ground infected tissue, 1 ml of PBS 1X buffer, and a small amount of carborundum. Leaves showing classic mosaic symptoms were ground and used to further inoculate the available leaves of maize plants in the third true leaf stage.
Three maize lines were used in this work. The resistant CI-RL1 line, from CIMMYT, was propagated in a greenhouse, and seeds were stored in a cold room. The susceptible line B73 was donated from the CIMMYT seed bank. Additionally, a F1 line was generated in the winter field in Vallarta, using CI-RL1 as the female and B73 as the pollen donor.
Leaves from mock-or virus-inoculated plants were surveyed by RT-PCR, for the presence of the CP cistron in B73 and F1 plants, in order to confirm viral infection after RNA extraction and quantification in a mix of four biological replicas. Overall, 12 pools of total RNA samples were generated: 2 for B73, 2 for RL, and 2 for F1 virus inoculates, plus 2 for each line for mock-inoculated plants. mRNA was isolated from each of the pools using Dynabeads, following the manufacturer's protocol. The concentration of each pool was measured again and the RIN of each of the pools was obtained. Samples were then shipped for RNA-seq analysis.

Experimental Design and Sequencing
Six plants of each line were grown in a greenhouse until they reached the third true leaf stage. All of the available leaves were inoculated with a mix of 500 mg of symptomatic ground tissue, 1X PBS buffer, and carborundum. Plants were observed daily and, at 17 dpi, leaves were collected and stored in an ultra-freezer at -70 °C. Total RNA was extracted from 100 mg of frozen tissue using TRIzol reagent (ThermoFisher, Waltham, MA, USA), according to the manufacturer's instructions. The concentration of the resulting total RNA was determined with Nanodrop 2000. Four biological replicates of each sample at the same concentration were mixed. mRNA was purified from the 12 total RNA pools using Dynabeads (Invitrogen, MA, USA), according to manufacturer's instructions. A total of 12 paired-end (2 × 150) RNA-Seq libraries (six treatments, two replicates per treatment) were prepared and sequenced at the Beijing Genomics Institute facility, using DNB-SEQ TM technology (BGI, CHN). The raw data are publicly available at the National Center for Biotechnology Information (NCBI), under BioProject accession PRJNA784537.

Mapping and Differential Expression Analysis
Trimming adaptors and cleaning of low-quality reads were conducted as described previously [93]. Mapping and expression quantification were conducted with Kallisto, as described by [94]. Reads that failed to map to the reference B73 transcriptome retrieved from monocots PLAZA 4.5 (https://bioinformatics.psb.ugent.be/plaza/versions/plaza_v4_5_monocots) were de novo assembled with Trinity under standard parameters [95]. Non-mapped reads were re-mapped to the de novo assembled transcripts and quantified as described earlier. The differential expression analysis was performed with Sleuth, as reported previously [96], comparing each infected maize line with their mock. Genes with a p-value of ≤0.01 were considered as differentially expressed genes (DEGs). Identification of expression pattern clusters (co-expressed transcript clusters) was carried out using the Clust software, under standard parameters with TPM normalization [97].

Gene Expression Validation
To validate the bioinformatic results of gene expression, we carried out RT-qPCR of 11 candidate genes ( Figure Supplementary 4), using the same RNA from sequencing. cDNA were synthesized using the RevertAid minus H enzyme. RT-qPCR reactions were performed in an CFX96 Real-time system (BioRad, Hercules, CA, USA) with Maxima SYBR Green qPCR/ROX qPCR Master Mix (2X) (Thermo Scientific, Linhus, Lithuania). Relative expression was calculated using the Ct values and normalized with β-tubulin as the housekeeping gene.

Conclusions
We performed RNAseq analysis of a resistant line (RL), a susceptible line (B73), and the F1 progeny of both lines. . GO enrichment analysis of cellular components led to the identification of putative candidate genes which were up-regulated in RL, coding for an integral component of the cellular membrane. We chose the gene annotated as an ABC transporter as a likely candidate for resistance. We also found two genes implied in longdistance movement-ZmPiezo and ZmPVIP-in maize. These three genes were validated using qRT-PCR.
Transcript expression patterns provided evidence of the diverse (and, hence, complex) behavior of the eIF4E homologs in maize, particularly the eEF1α factors. Furthermore, two new maize gene models-ZmeIF4E and ZmeEF1α-were identified as possibly being involved in the SCMV interaction.
More work is needed to understand how the genes identified in this work-ZmeIF4E, Zm-elfa, ZmPiezo, ZmPVIP and ABC transporter-are involved in maize potyvirus resistance.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Table S1: eIFs and eEFs annotated in the MaizeGDB database.; Table S2: List of the total of resistance candidate genes searched for in the expression pattern clusters.; Table S3: Candidate genes found in group 1 (blue) and group 2 (red) of the expression pattern clusters.; Figure S4: Validation of candidate genes by RT-qPCR.
Funding: G.R. and P.V. were supported by Consejo Nacional de Ciencia y Tecnologia postgraduate scholarships (CVU/Becario: 420714/262948 and 1024179/, respectively). This research was funded through PlanTecc and Cinvestav funds, in addition to personal funding by L.S. Data Availability Statement: