Expression patterns of NbrgsCaM family genes in Nicotiana benthamiana and their potential roles in development and stress responses

rgsCaM has been reported as a calmodulin-like (CML) factor induced by viral infection in Nicotiana. There are three CMLs that belong to the rgsCaM family in Arabidopsis thaliana. In this study, we found a total of 5 NbrgsCaM coding sequences in N. benthamiana genome. We analyzed transcription patterns of NbrgsCaMs in transgenic plants expressing a β-glucuronidase (GUS) under the promoter of NbrgsCaMs by histochemistry staining and RT-qPCR. Similar to their Arabidopsis homologs, most NbrgsCaMs have an overlapping but distinct expression pattern in response to developmental and environmental changes. Specifically, the NbrgsCaM4 promoter exhibited robust activity and showed distinct regulatory response to viral infection, developmental stages and other abiotic stimuli. Overall, these findings provide clues for further understanding of the NbrgsCaM family genes in regulating plant growth and development under biotic stress and environmental stimulation.

Analysis of the promoters of rgsCaM family. Our analysis (by Softberry TSSP program online prediction, http://linux1.softberry.com/ 22 ) of the regions between the stop codon of the previous gene and the start codon of AtCML37, 38 and 39, which are putative promoters named as AtrgsCaMps or AtCMLps, revealed promoter and enhancer motifs (listed in Supplementary Data S4) that relate to development, phytohormonal regulation and environmental stresses. The correlation between transcription factors and the predicted transcription factor binding sites/RegSites can be found in the Softberry RegSite database (http://www.softberry.com/berry. phtml?topic=regsitelist) 23 . For clarity, we did not list the candidate transcription factors for each promoter/ enhancer motifs. Instead, we summarized the transcription factors for these predicted motifs (Supplementary  Table S1), consistent with the reported involvement of AtrgsCaM in development and stress responses 5,6,8,24,25 .
The promoters of NbrgsCaMs also contain multiple regulatory elements (Supplementary Data S5), many of which are the same type of motifs that exist across the rgsCaM family. The transcription factors that are predicted for the recognition of these regulatory elements are listed, together with their specific functions (Supplementary  Table S2). According to the description of these transcription factors, NbrgsCaMps should be related to multiple developmental, environmental and plant hormonal regulations, and with overlapping but distinct stimuli response patterns among the NbrgsCaMps (Supplementary Table S3). Thus, combining the promoter analysis of AtrgsCaMps (Supplementary Table S1) and the reported function of AtrgsCaMs 5,6,8,24,25 , it can be deduced that rgsCaMs are not only homologous in coding sequences, but also respond to similar types of environmental and developmental stimuli in Nicotiana and Arabidopsis. However, the promoters of rgsCaMs are rather conserved in regulatory motifs but not in sequence, and the enhancer motifs do not hold the same positions (Supplementary Data S4 and S5), suggesting that the functionally conserved rgsCaMs evolved subtle and distinct regulatory niches to carry out precise regulatory work. β-glucuronidase (GUS) reporter analysis of the promoters of rgsCaM family genes. To analyze the expression profile of rgsCaMs, we constructed chimeric rgsCaM-promoter::GUS (rgsCaMp::GUS) reporters. GUS staining of the transgenic N. benthamiana with rgsCaMp::GUS reporters was carried out at various developmental stages and under specific stress treatments.
The transgene of rgsCaMp::GUS reporters (NbrgsCaMp1::GUS, NbrgsCaMp3::GUS, NbrgsCaMp4::GUS, NbrgsCaMp5::GUS and NbrgsCaMp7::GUS) afflicted no impact on the growth of N. benthamiana, indicating that our reporter system didn't intervene with the physiology of transgenic plants. The GUS staining in these transgenic plants can actually reflect the endogenous promoter activity of these NbrgsCaMps. In seedlings of the N. benthamiana transgenic lines, GUS expression was detected in NbrgsCaMp1::GUS, NbrgsCaMp3::GUS, NbrgsCaMp4::GUS and NbrgsCaMp5::GUS, while NbrgsCaMp7::GUS showed no observable staining at all (Fig. 2a). Among the transgenic plants with GUS expression, NbrgsCaMp3::GUS and NbrgsCaMp4::GUS showed darker GUS staining than NbrgsCaMp1::GUS and NbrgsCaMp5::GUS (Fig. 2a). The GUS transcription level was further measured by reverse transcription quantitative polymerase chain reaction (RT-qPCR). The RT-qPCR result affirmed that the GUS staining indeed reflected the transcription level that is the indicator of promoter activity (Fig. 2b). As the seedlings grew older, we found that more GUS staining accumulated in the root of NbrgsCaMp4::GUS transgenic N. benthamiana plants (Fig. 2c), revealing that NbrgsCaMp4 activity is developmental stage-and tissue-associated. To systemically analyze the NbrgsCaMps activity, we carried out GUS staining assays in various plant organs. The results showed that, in mature leaves, NbrgsCaMp3 and NbrgsCaMp4 were the most robust promoters among all of the NbrgsCaMps, while NbrgsCaMp1 and NbrgsCaMp5 were much weaker, whereas the GUS staining in NbrgsCaMp7::GUS was almost undetectable. Likewise, in flowers, the promoter activity was robust for NbrgsCaMp3, NbrgsCaMp4 and NbrgsCaMp5 with heavily stained anthers and sepals, whereas in anthers, NbrgsCaMp1::GUS showed weak staining, and NbrgsCaMp7::GUS had no detectable GUS staining. Moreover, the promoter activity was obvious for NbrgsCaMp4, modest for NbrgsCaMp3 and NbrgsCaMp5, marginal for NbrgsCaMp1, and undetectable for NbrgsCaMp7 in roots (Fig. 3a). Overall, NbrgsCaMp4 was the most active one among the NbrgsCaMps (Fig. 3a), with the GUS expression level in organs increased in the order of leaf, flower and root (Fig. 3b). The other NbrgsCaM promoters exhibited less activity except in the leaves and flowers of NbrgsCaMp3::GUS when compared to their counterparts of NbrgsCaMp4::GUS.

Analysis of NbrgsCaMps activity under salt and peG treatment. To investigate whether NbrgsCaMps
respond to environmental stresses, we applied salt and PEG-simulated drought treatment to the transgenic seedlings. After transferring the seedlings to media containing additional salt or PEG for 5 days, GUS staining was carried out to detect the NbrgsCaMps activity. Elevated GUS staining was detected in both salt and PEG treatment for NbrgsCaMp4::GUS compared to the untreated ones. However, NbrgsCaMp5::GUS responded mainly to salt, and NbrgsCaMp1::GUS responded mainly to PEG, while the other samples showed no detectable response to the application of either salt or PEG (Fig. 4).

Analysis of rgsCaMps activity after viral infection.
To investigate the response of NbrgsCaMps to viral infection, we inoculated N. benthamiana plants with DNA A of Tomato yellow leaf curl China virus (TYLCCNV A), TYLCCNV A together with its β satellite (TYLCCNV A + β), Potato virus X (PVX), Tobacco mosaic virus (TMV), or inoculation buffer, which was used as a negative control. Then we compared the GUS staining of the systemic leaves. Because NbrgsCaMps activities are developmental stage-related, we carefully chose the leaves at the same developmental stage for comparison. TYLCCNV A and TYLCCNV A + β triggered induction of GUS expression in symptomatic systemic leaves of NbrgsCaMp3::GUS, NbrgsCaMp4::GUS and NbrgsCaMp5::GUS at 7 days post infection (dpi) (Fig. 5a). The induction difference was observed between NbrgsCaMp4::GUS infected by TYLCCNV A + β and by TYLCCNV A, and the TYLCCNV A + β infection induced heavier GUS staining (Fig. 5a), which was also confirmed by RT-qPCR of the expression of NbrgsCaM4 in wild type N. benthamiana plants at 12 dpi (Fig. 5e). However, we failed to observe significant elevation of GUS expression in NbrgsCaMp4::GUS infected by TYLCCNV A + β comparing to TYLCCNV A and buffer only inoculation at 12 dpi (Fig. 5d). There was a difference in fold change of GUS expression between TYLCCNV A + β and buffer only inoculation, but was not as high as the previously reported case 10 . This might be due to the less sensitivity of transgene reporter and large variation between biological repeats. Nonetheless, RT-qPCR of the systemic leaves at 5 dpi (the time when viral symptoms first appear) showed that TYLCCNV A + β and TYLCCNV A triggered no difference in the expression of either GUS or NbrgsCaM4 in NbrgsCaMp4::GUS transgenic or wild type N. benthamiana plants respectively (Fig. 5b,c). Additional analysis of the systemic leaves of NbrgsCaMp4::GUS infected by PVX or TMV showed that both viruses triggered induction of GUS expression, as evidenced by the GUS staining ( Fig. 6a,b) and RT-qPCR results (Fig. 6c,d). Additionally, after inoculation with PVX.βC1 (a PVX-based viral vector for overexpression of geminivirus viral factor βC1), transient overexpression of βC1 induced higher GUS expression than by inoculation with PVX alone (Fig. 6a,c), resembling the effect of DNAβ in TYLCCNV infection (Fig. 5a,e).
Our results confirmed that inoculation with the viruses listed above (TYLCCNV A, TYLCCNV A + β, PVX, TMV, PVX.βC1), all induced GUS staining in symptomatic systemic leaves of NbrgsCaMp4::GUS. In both the TYLCCNV A + β-infected plants and the PVX.βC1-infected plants, the existence of βC1 triggered higher NbrgsCaMp4 activity and NbrgsCaM4 expressions after the initial display of symptoms. The exacerbating effect of βC1 manifested its function as a viral pathogenicity factor. However, the TYLCCNV A + β inoculated plants failed to display increase of GUS induction compared to TYLCCNV A when the early symptoms displayed at 5 dpi. It is still unclear whether the presence of βC1 induced the activation of NbrgsCaMp4 directly or through accumulated pathogenicity of the virus. on the activation of NbrgsCaMp4, we directly measured the transcription of NbrgsCaM4 by RT-qPCR. The transcripts level of NbrgsCaM4 was higher in βC1 transgenic N. benthamiana than in WT (Fig. 7a,b). As the βC1 transgenic N. benthamiana takes much longer time to grow (usually more than a year to reach the flowering stage) than the WT (50 days to reach the flowering stage), we tried to rule out the influence of leaf age on NbrgsCaM4 expression. The results showed that, after the seedling stage, the activity of NbrgsCaMp4 increased as the plant aged (Fig. 7c,d). Similarly, further analysis of both the NbrgsCaM4 expression in WT and the GUS expression in NbrgsCaMp4::GUS showed that the activity of NbrgsCaMp4 increased as the leaves aged even in the same plant ( Fig. 7e-g). Thus, the influence of age on the activity of NbrgsCaMp4 between the βC1 transgenic and WT N. benthamiana becomes more important. To clarify this factor, we further tested the NbrgsCaM4 expression in seedlings of the βC1 transgenic and WT N. benthamiana at 10 days after germination (Fig. 7h). The RT-qPCR www.nature.com/scientificreports www.nature.com/scientificreports/ results showed that the activity of NbrgsCaMp4 is higher in βC1 transgenic N. benthamiana (Fig. 7i). However, the βC1 transgenic N. benthamiana is deformed with needle-like leaves (Fig. 7a). This type of needle-like leaves mostly are composed of veins due to the pathogenicity of viral factor βC1 ( Supplementary Fig. S2). As GUS staining accumulated mainly in veins of the NbrgsCaMp::GUS (Figs. 2a, 3a, 4 and 5a), it is difficult to exclude the influence of tissue expression bias to draw a conclusion that the up-regulation of NbrgsCaMp4 activity is directly generated by βC1 in βC1 transgenic N. benthamiana.
Analysis of NbrgsCaMps activity under wounding treatment. Viral symptoms progressed as the infection persisted, and necrotic spots and yellowing usually accumulated in the infected leaves as the symptoms aggravated. For N. benthamiana, the infection by TYLCCNV A + β can induce higher activity of NbrgsCaMp4 than the infection by TYLCCNV A after prolonged infection, but not at the early stage, i.e. at 12 dpi (Fig. 5e) instead of at 5 dpi (Fig. 5c). This made us wonder whether the accumulated damaging effect of TYLCCNV combined with βC1 is the direct trigger of NbrgsCaMp4 activity, rather than the expression of βC1 itself. To investigate the damaging effect on NbrgsCaMp4 activity, we infiltrated the inoculation buffer to the leaves of NbrgsCaMp4::GUS transgenic N. benthamiana plants to impose mechanical damage without introducing viral infection. DAB staining carried out at 1 h after infiltration revealed dark spots around the injection sites, indicating the presence of damages (Fig. 7j). RT-qPCR results showed that the transcription of GUS in buffer-only-infiltrated leaves elevated up to 200 fold, compared to that of the untreated control at 1 h post infiltration (Fig. 7l). This transcription elevation subsided quickly. At 3 h post infiltration, it was still about 10 fold higher than the control but not that dramatic (Fig. 7l). On the other hand, in the GUS staining results, the leaves were stained more heavily at 3 h than at 1 h, a little postponed, which might be due to the accumulation of GUS expression (Fig. 7k). These results indicated that NbrgsCaMp4 responded instantly to damage stress. As viral infection can afflict damages to plants, the up-regulation of NbrgsCaMp4 activity can be attributed to the side effect of viral symptoms.

Discussion
NbrgsCaMs form a unique branch of CML subfamily with multiple members. While rgsCaM was first discovered in Nicotiana 9 , and had been assigned complicated functions in viral responses 7,[9][10][11][12] , it has been considered the only rgsCaM in Nicotiana for about 20 years. Its close homologs have been left in the shadows until this study, in which our analysis revealed 5 coding genes and 2 possible pseudogenes in N. benthamiana (Supplementary Data S1-S3). We named these coding genes and pseudogenes NbrgsCaM1-7, sequentially according to the scaffold serial numbers they reside in. NbrgsCaM2 and NbrgsCaM6 are counted as pseudogenes because their ORFs show no similarity to other proteins except the CaM family, and even their longest ORFs contain only an incomplete CaM function motif. Coding sequences of NbrgsCaM1, 3, 4, 5 and 7 share identities ranging from 70.4% to 93.3% (amino acid sequences identity ranging from 60.2% to 83.2%) (Supplementary Data S2 and S3). Like AtCML37, 38 and 39, the AtrgsCaMs in Arabidopsis 4 , NbrgsCaMs form a unique subclass of CMLs in N. benthamiana (Fig. 1), with close similarities to each other. In the phylogenetic tree, AtCML40 and AtCML41 coexist in the same subclass with the reported rgsCaMs (Fig. 1c). These two genes have not yet been thoroughly studied, as had the AtCML37, 38 and 39 5,6,8,24,25 . Further studies of them might be able to provide more clues to the role of rgsCaMs in plant development and stress responses. www.nature.com/scientificreports www.nature.com/scientificreports/ NbrgsCaMs display differential expression during development and under environmental stresses. There are about 50 CMLs in the Arabidopsis genome, answering to Ca 2+ fluctuations generated through nearly all environmental, hormonal and developmental stimuli 3 . As a close subfamily of CMLs, not only the resemblance in amino acid sequences, but also the existence of identical enhancer motifs in the promoters of AtCML37, 38 and 39 correlate with their similar but distinct responses to environmental and developmental stimuli (Supplementary Table S1) 5,6,8,24,25 . Similarly, our analysis revealed that NbrgsCaMps are enriched with developmental and stress specific enhancer elements, many of which also exist in the promoters of AtrgsCaMs, and each NbrgsCaMp has a specific combination of enhancer motifs (Supplementary Table S2 and S3). The disclosure of the amino acid sequences and enhancer elements of NbrgsCaMs indicates that they possess overlapping but non-identical regulatory functions in N. benthamiana. NbrgsCaMp::GUS reporter analysis revealed tissue-and developmental stage-specific promoter activities of NbrgsCaM1, 3, 4 and 5 (Fig. 3), corroborating that NbrgsCaMs are important developmental regulatory factors. Furthermore, salt and PEG treatment induced elevated GUS expression in most of the NbrgsCaMp::GUS transgenic N. benthamiana (Fig. 4), demonstrating that NbrgsCaMps respond to environmental stresses as well. Each of these NbrgsCaMps drives a specific GUS expression pattern. In vegetative tissue, the GUS staining is most pronounced in veins of leaves and roots (Figs. 2a, 3a, 4 and 5a), corresponding to the vascular and root development-related promoter motifs: ATHB-2, AtMyb77, RAV1 and Root-specific nuclear factor enhancer elements (Supplementary Table S2). Among the NbrgsCaMps, NbrgsCaMp3 and 4 have the strongest GUS staining in both seedling and mature leaves (Figs. 2 and 3); NbrgsCaMp4 and 5 responded more to salt stress than the other NbrgsCaMps did (Fig. 4); while NbrgsCaMp4 is the strongest NbrgsCaMp in roots (Fig. 3a). All in all, NbrgsCaMp4 is the most highly active promoter among the NbrgsCaMps during vegetative growth, and responds actively to salt and PEG stress. In flower, NbrgsCaMp4 is also one of the strongest NbrgsCaMps (Fig. 3). Thus, there is no wonder that NbrgsCaM4 was the first rgsCaM discovered in N. benthamiana due to its predominant expression level 13 . On the other hand, although we failed to detect promoter activity in NbrgsCaMp7 by GUS staining in the NbrgsCaMp7::GUS transgenic lines, the expression of other NbrgsCaMs are not to be neglected according to their promoter activities that were shown clearly by the GUS staining and RT-qPCR analyses, and undoubtedly they play important roles, considering that the activities of NbrgsCaMp1, 3, and 5 are robust during development and under stress treatments (Figs. [2][3][4][5]. The evidence of NbrgsCaMs expression can also be found in the RNA-seq data provided by the Sol Genomics Network (https://solgenomics.net/jbrowse_solgenomics/) 18,26 . The RNA_seq reads count can viewed by typing in the scaffold location of specific genes on JBrowse, the Sol Genomics Network. Despite the presence of gaps and assembly incompleteness in genomic region of some NbrgsCaMs, it is clear that RNA_seq reads for NbrgsCaM1, 3, 4 and 5 are abundant 18,26 . The RNA_seq reads count for NbrgsCaM4 is the highest, and for NbrgsCaM7 is the lowest, which is only about 1/10 of that of the NbrgsCaM4.

tYLccnV infection induces the activity of NbrgsCaMp3 and NbrgsCaMp5, in addition to
NbrgsCaMp4. NbrgsCaM4 was the first discovered NbrgsCaM 13 . The namesake of "rgs" came from its role as a regulator of gene silencing in virus-plant interaction 9 . "rgs" is by far the pivotal role studied for Nicotiana rgs-CaM, despite the discovery of developmental and stress-related functions for AtrgsCaMs in Arabidopsis [5][6][7][8][9][10]24,25 . In this study, we investigated NbrgsCaMps responses to viral infection. Increased induction of GUS expression was detected in the systemic leaves of NbrgsCaMp::GUS for not only NbrgsCaMp4, but also NbrgsCaMp3 and 5 after inoculation with TYLCCNV A and TYLCCNV A + β, compared to those of the untreated plants (Fig. 5a). βC1 induced the expression of NbrgsCaM4 through a damaging side effect of its virulence. We detected elevated NbrgsCaM4 expression in TYLCCNV A + β inoculated wild type N. benthamiana, when compared to those inoculated with TYLCCNV A at 12 dpi (Fig. 5e), similar to a previous report 10 . But quite unexpectedly, we found that TYLCCNV A and TYLCCNV A + β treatment induced equivalent GUS or NbrgsCaM4 expression levels in systemic leaves of NbrgsCaM4::GUS and wild type N. benthamiana at 5 dpi (Fig. 5b,c). The time point of 5 dpi is when viral symptoms shown up in the TYLCCNV A + β inoculated systemic leaves. The initial viral symptoms of TYLCCNV A + β infection appeared in our observations as wrinkled curly leaves and bulging veins at 5dpi, demonstrating the presence of βC1 in the leaves that we analyzed at this time point. This suggested that βC1 possibly is not the direct factor for NbrgsCaMp4 activation at the early stage of viral infection. Furthermore, the GUS expression levels were induced by RNA viruses, PVX and TMV (Fig. 6), similar to the results reported by Chung et al. 13 , suggesting that the induction of NbrgsCaMp4 is rather a general response to viral infection, indiscriminate of DNA or RNA viruses. We have used untreated instead of mock-inoculated N. benthamiana plants as controls for TMV infection. This is the only place where the untreated N. benthamiana plants were used instead of mock-inoculated ones. As the damage induced activity of NbrgsCaMp4 subsides quickly (Fig. 7l), and no induced upregulation of GUS expression has been detected in the mock-inoculated NbrgsCaMp4::GUS transgenic N. benthamiana plants after 3 dpi, we consider the untreated plants to be as sufficient a control as the mock-inoculated ones in TMV infection. In addition, the presence of βC1 in TYLCCNV A + β and PVX.βC1 both induced higher activity of NbrgsCaM4 promoter (Figs. 5 and 6), suggesting that βC1 can promote the expression of NbrgsCaM4 together with its natural master virus TYLCCNV A or the artificial viral vector PVX.
The application of βC1 together with its master viral DNA -TYLCCNV A, or with an RNA viral vector -PVX, introduced other viral factors that probably interfered with NbrgsCaM4 expression. To single out βC1 for further analysis, we directly analyzed NbrgsCaM4 transcription in transgenic N. benthamiana for βC1 overexpression. Although in both seedlings and mature plants, the expression level of NbrgsCaM4 was higher in βC1 transgenic plants than in WT (Fig. 7a,b,h,i), we still cannot solely attribute the induction of NbrgsCaM4 to βC1 directly, as we can't rule out the impact of morphology changes and extended vegetative growth stage of the βC1 transgenic plants on NbrgsCaM4 expression. Our analysis of the expression levels of NbrgsCaM4 and GUS in WT (2020) 10:9652 | https://doi.org/10.1038/s41598-020-66670-x www.nature.com/scientificreports www.nature.com/scientificreports/ and NbrgsCaMp4::GUS transgenic N. benthamiana plants respectively, provided evidence that the NbrgsCaM4p activity increased greatly according to the advance of leaf aging in mature plants (Fig. 7c to g). Furthermore, the GUS staining accumulated in veins of NbrgsCaMp4::GUS (Figs. 2a, 3a, 4 and 5a), corroborating with the presence of vascular-specific enhancer element in NbrgsCaMp4 (Supplementary Table S2). As the deformed leaves of βC1 transgenic N. benthamiana are composed mostly of vascular tissue ( Supplementary Fig. 2), and take a very long time to grow, it is hard to justify whether the elevated NbrgsCaMp4 activity came from the skewed development caused by βC1 transgene or from the βC1 factor directly. Thus, it is more reliable to analyze the induction of NbrgsCaM4p through TYLCCNV A + β treatment than in βC1 transgenic plants.
TYLCCNV A is a mild virus that induces almost no symptoms to N. benthamiana 27,28 . On the other hand, βC1 is a pathogenic factor which is responsible for the severe viral symptoms generated by TYLCCNV A + β, such as curly leaves 29 . Thus, it was considered to be the trigger of many physiological changes in the infected plants. The fact that the TYLCCNV A + β and TYLCCNV A triggered divergence in NbrgsCaM4p activity happened several days after the appearance of viral symptoms is largely the reflection of the side effect of βC1 virulence, consistent with the appearance of yellowing in severely infected plants, the signature of damages and aging, which happened at the late stage of infection. It is also worth noticing that the TYLCCNV A + β treatment is not acute in induction of the NbrgsCaMp4 activity, compared to the instant and dramatic elevation of NbrgsCaMp4 activity triggered by wounding (Fig. 7j to l).
The elevation of GUS and NbrgsCaM4 expression during aging and wounding is dramatic (Fig. 7c to g, and 7j to 7 l). But, the increase of their expression during viral infection is rather moderate (Fig. 5). Thus, compared to that during aging and wounding, the promoter activity of NbrgsCaMp4 is far less robust under viral infection. As a regulatory factor, efficient responses are necessary for cascading amplification of signals to cope with environmental and developmental fluctuations of the surroundings. The moderate responses of NbrgsCaMp4 to viral infection suggest that the inflictions from viral factors are not the major situations that NbrgsCaM4 evolved to cope with, which means that probably βC1 induced the expression of NbrgsCaM4 indirectly through damages to plants by aggravating viral symptoms. RNA-interference (RNAi) has been reported as a surveillance system that protects the shoot tips from viral infection 30,31 . Since only low levels of NbrgsCaM4 expression have been detected in the young leaves, especially for the very young leaves near shoot tips, it is less likely that NbrgsCaM4 can effectively suppress plant RNAi as viruses propagate (Fig. 7c to f In summary, the NbrgsCaMs form a distinct class of CMLs (Fig. 1c). There are more than one rgsCaM in both Arabidopsis and N. benthamiana (Supplementary Data S3, Fig. 1) 4,8 . They respond to developmental and environmental changes, particularly salt, drought and wounding stresses, via elevated expression to meet the need of timely regulation, and have overlapping but still distinct expression atlases which have been demonstrated by their promoter activities (Figs. 2-7). NbrgsCaM4 is the most robustly expressed rgsCaM in N. benthamiana according to the promoter activity analysis (Fig. 2-5) and RNA-seq data from the Sol Genomics Network 18,26 . Fitting its role in regulation, NbrgsCaMp4 responds to diverse stresses, including viral infection, though its response to viral infection is mild and probably indirect through damages owing to a side effect of viral symptoms (Figs. 6 and 7). Thus, rather than being induced by a specific viral factor, NbrgsCaM4 is a member of a CML subfamily that response mostly to general developmental stages and stresses. Despite the importance of NbrgsCaM4, other NbrgsCaMs also respond actively to environmental stimuli, with overlapping expression patterns, and are probably also overlapping in functions with respect to their close homology in protein sequences. The findings of this study are helpful in characterizing not only the expression patterns, but also the relative expression strength of rgsCaM genes, being the first step towards a future understanding of the rgsCaM family's multiple functions.

Materials and Methods
Plant materials and growth conditions. Wild type (WT) and 35 S::βC1 transgenic (generated by Qiuying Yang according to the method described before 29 ) N. benthamiana seeds were surface-sterilized with 75% ethanol and 25% bleach for 1 min and 3 min respectively, and then washed three times with sterile water. Sterile seeds were plated on Murashige and Skoog (MS) medium plus 2.0% sucrose and 0.5% phytagel. Plated seeds were placed in a greenhouse set at 24 °C, 16-h-light/8-h-dark photoperiod for germination and growth. For culturing of mature plants, seedlings were transferred to soil after 5 days on plates and placed in the same greenhouse as above. Plants were watered as required and supplemented every other week with fertilizer. Seedlings and tissues were harvested at various time points for GUS staining and RT-qPCR analysis.
construction of alignments and trees. Sequences of CaM and CML proteins were downloaded from the Arabidopsis Information Resource (TAIR) (http://www.arabidopsis.org) and the Sol Genomics Network (https:// solgenomics.net/organism/Nicotiana_benthamiana/genome) and subjected to phylogenetic analysis. All of the www.nature.com/scientificreports www.nature.com/scientificreports/ A. thaliana CaMs and CMLs have been listed in the previous publication 4 , except for CML51 (At1g73440), which was added in this study. While the NbCaMs and NbCMLs were obtained by blasting the Sol Genomics Network data with A. thaliana CaMs. Alignments were constructed using the alignment mode of ClustalW in MEGA X 32 . Note: calcineurin B-like calcium sensor proteins and calmodulin-domain protein kinases are not included as CMLs in the analyses. Multiple alignment of protein sequences was carried out with the following alignment parameters: gap opening penalty of 10, gap extension penalty of 0.2, negative matrix off and delay divergent cutoff of 30%. Protein trees were constructed using the maximum likelihood method with bootstrap of 500 embedded in the MEGA X software. Parameters for multiple alignment of nucleotide sequences were: gap opening penalty of 15, gap extension penalty of 6.66, DNA weight matrix IUB, transition weight 0.5, negative matrix off and delay divergent cutoff of 30%.
Generation of nbrgscaMp::GUS transgenic n. benthamiana. The promoter sequences of NbrgsCaMs have not been reported yet. We took 1kbp before the start codon of these genes as the putative promoters, as most of the NbrgsCaMs promoter elements locate within this region according to the online prediction results in this study. We cloned and constructed the promoter sequences into pBI101.GUS to generate pBI101. NbrgsCaMpn::GUS (n = 1, 3, 4, 5, 7, order of the NbrgsCaM homologs). Primers used in the construction are listed (Supplementary Table S4). The plant transgene was done by the Bio-run company (http://www.biorun. com/). More than 10 T0 positive transgenic lines were tested for each NbrgsCaMpn::GUS transgene, and homozygous T1 lines were obtained through self-fertilization of T0 plants.
Stress treatments: salt and peG. Salt and drought-simulation (using PEG) treatments of the NbrgsCaMp::GUS transgenic N. benthamiana plants were conducted using 5-day-old seedlings. For each stress treatment, seedlings were carefully removed from the MS plates and transferred to plates supplemented with stress reagents, and grown for an additional 5 days. Salt stress plates were simply the MS plates supplemented with 200 mM NaCl; drought simulation plates were prepared by equilibrating the MS plates with 20% PEG8000 solution (filter sterilized) overnight 33,34 . At least three independent transgenic lines of each NbrgsCaMp::GUS were analyzed. Samples were collected after the 5 days salt or PEG treatments for GUS staining and RT-qPCR analysis.

Stress treatments: viral infection and wounding. For viral infection, N. benthamiana was grown for
4 weeks after transferring to soil to get ready for viral inoculation. Leaves were agro-inoculated with TYLCCNV A 29 , TYLCCNV A + β 29 , PVX 35,36 , PVX.βC1 37 , TMV 38 or mock-inoculated with inoculation buffer as described 39 . Systemic leaves from the infected plants, and equivalent leaves from the mock-inoculated ones were harvested for GUS staining and RT-qPCR analysis.
Infiltration of inoculation buffer into the mock-inoculated local leaves can cause mechanical wounding. Local leaves were harvested at specific time points for 3,3'-diaminobenzidine (DAB) staining to detect hydrogen peroxide 40 , the signal of damage generated after wounding.
RnA extraction and Rt-qpcR analysis. Total RNA was isolated using TRIzol method (Invitrogen). RNA concentration and quality were determined by spectrophotometry and gel electrophoresis. For RT-qPCR analysis, total RNA was treated with DNase I (Takara) and reverse transcribed according to the manufacturer's instructions (EasyScript cDNA Synthesis SuperMix kit, TransGen Biotech). Specific primer pairs for NbrgsCaM4, GUS and GAPDH (an internal control) were listed in Supplementary Table S4. qPCR was performed using Roche LightCycler 96 with TransStart Green qPCR SuperMix (TransGen Biotech). Primer pairs were validated by cDNA template titration to ensure similar amplification kinetics and a single melting point of quantitative PCR products. Each experiment was performed in triplicate and repeated three times with different biological samples, and the results were analyzed with software supplied by the manufacturer. We used comparative CT method to determine the relative expression level of target gene expression 41 . Levels of the housekeeping gene GAPDH were used to calculate changes (n-fold) by comparing mean threshold cycle values. P value < 0.05 is used to delimit statistical significance. Error bars indicate S.D. To avoid genomic DNA contamination, a reaction lacking reverse transcriptase was performed in parallel for each sample.
Histochemical assays: GUS and DAB staining. For GUS staining, leaves were immersed in GUS staining solution (GUS staining kit, HUAYUEYANG biotechnology co., LTD.) for 12 h at 37 °C in darkness, and then washed with 70% ethanol several times to remove background 42 . Samples and controls that were to be compared together were always stained in the same batch to eliminate variations caused by altering of conditions during the experiments. Stained samples were observed with Olympus SZX16 microscope (10 X amplification) and photographed by digital camera (Olympus DP72).
For DAB (3,3'-diaminobenzidine) staining, agro-infiltrated leaves were incubated in 1.0 mg/mL DAB-HCl solution in the dark overnight, then destained by boiling in 95% ethanol for 5 min. Dark brown precipitates on the leaves indicate detection of hydrogen peroxide generated after wounding 43 . Section of the plant leaves. Leaves of the WT and 35 S::βC1 transgenic N. benthamiana plants were fixed in FAA fixation buffer (containing 50% EtOH, 5% HAc and 3.7% formaldehyde) and sent to the SanShu Biotech Company (http://www.sanshubio.com) for resin embedded dissection and observation with high-resolution light microscopy according to the protocol 44 .