Introduction

Cortical microtubules (MTs), which consist of heterodimers of α-tubulin (TUA) and β-tubulin (TUB) proteins, are essential to the plant cell morphogenesis and play a key role in guiding the deposition of cellulose microfibrils during plant cell wall formation1,2. The amino acid sequences of the TUA and TUB subunits show about 88% homology in animals3,4,5, plants6, protists7,8 and fungi9. The TUA and TUB genes in plants have been extensively investigated; for example, Arabidopsis thaliana harbors six TUA genes that encode four distinct proteins10,11 and at least nine TUB genes that encode nine proteins12. On the other hand, cotton (Gossypium hirsutum) contains five TUA genes13 and nineteen TUB genes14; rice (Oryza sativa) has four TUA genes15,16 and eight TUB genes17 and Populus trchocarpa possesses eight TUA and twenty TUB genes18.

The expression pattern of plant tubulin has also been studied in various species. The Arabidopsis AtTUA1 gene is only expressed in flowers and AtTUA2AtTUA5 are transcribed in leaves, roots and flowers10,19. Of the nine of AtTUB genes, the transcripts of AtTUB5 and AtTUB6 preferentially accumulate in leaves and petioles12, whereas the AtTUB1 transcript is primarily expressed in the roots20 and the other seven TUBs are expressed in vegetative tissues. Rice OsTUB8 is predominantly expressed in flowers and other seven OsTUBs are differentially expressed during development17. The transcript levels of five cotton TUAs are much higher in fibers than that in various other tissues, including pollen13 and 9 of 19 GhTUB genes are preferentially expressed in cotton fiber cells14. Populus PtTUA1, PtTUA5, PtTUA7, PtTUB5/6, PtTUB7, PtTUB8, PtTUB19 and PtTUB20 have the highest transcript levels in pollen, whereas the other PtTUAs and PtTUBs are upregulated in the xylem18. Functionally distinct microtubule subtypes are generated in cells through the expression of different tubulin isotypes and through post-translational modifications (PTMs). In animals, tubulins have different homologs that undergo various PTMs such as tyrosination/detyrosination, acetylation, polyglutamylation and polyglycylation, which in turn lead to the appearance of various tubulin isoforms and classes of MTs21,22,23,24,25,26,27. In plants, a large number of tubulin isotypes have been isolated from different species, whereas investigations on PTMs in plant tubulin are limited.

Willow (S. arbutifolia) is a fast-growing tree and is cold-resistant, thus rendering it an ideal material for biomass production research. The interest of using willow for biomass production is growing, thereby resulting in increased pressure on breeding high yielding and resistant clones that are adapted to different environments28. MTs play central roles in several basic processes of eukaryotic cells, which include cell division, cell motility, intracellular transport and cell elongation. The long-term goal of our investigations is to characterize the mechanisms involved in the genetic control of MT function in woody plants. In the present study, we report the cloning, sequencing and analysis of gene structure, phylogenetic diversity and expression patterns of eight TUA genes and twenty TUB genes in S. arbutifolia. We also analyzed the amino acid residues at the C-terminal region of the TUA and TUB proteins and our results have prompted us to speculate that the other PTMs are related to the new C-terminal amino acid residues in Salix.

Materials and Methods

Identification of α- and β-tubulin genes in Salix

The whole-genome sequence data of Salix. suchowensis were used for the identification of α- and β-tubulin genes, including both of DNA and CDS, via reciprocal BLAST analysis using protein sequence of 20 Populus and 15 A. thaliana tubulin genes29. Salix homologs to the 20 Populus tubulin genes were identified by using BLASTP, with the e-value cut-off set at 1-E03. The same protocol was performed for the detection of willow homologs by using the 15 A. thaliana tubulin genes.

DNA cloning and sequencing

Total RNA was extracted and treated with RNase-free Dnase (Promega, Madison, USA) to remove contaminating DNA. Purification of first-strand cDNA was conducted following the protocol of Lu et al.30. Primers (Supplementary file 1: Table S1) were designed according to the sequences of the α- and β-tubulin genes, which were obtained by analyzing the Salix genome. PCR was performed as follows: 94 °C for 2 min, followed by 30 cycles of 94 °C for 30 s, 56 °C for 45 s and 72 °C for 2 min. The PCR products were cloned into the pMD18-T vector (Takara, Japan, http://www.takara.com.cn) and sequenced. The 28 cDNAs (8 TUAs and 20 TUBs) from Salix were designated as SaTUA1–SaTUA8 and SaTUB1–SaTUB20, respectively.

Real-time PCR Analysis

Stem developing phloem, full expanded leaves, stem developing xylem, shoot tips (1.0 cm–1.5 cm from the top of the plant) and inflorescence were obtained from three 1-year-old S. arbutifolia, which were growing at the Baishilazi National Nature Reserve of China. Real-time PCRs were conducted using the SYBR Green Perfect (Takara, Dalian, China) and StepOnePlusTM System (Applied Biosystems). All of the PCR products were sequenced and the dissociation curve was analyzed to verify amplification specificity31. The purified PCR products were employed to generate a standard curve to establish the quantitative correlation between the CT values and the transcript copy numbers31. Each qRT-PCR reaction was repeated at least three times and each standard curve comprised at least 5 points. The transcript levels of the samples did not significantly vary (P-value ≥0.95) from the calculated average of the biological replicates.

Sequence analysis

Phylogenetic analysis was performed using the software, MEGA 532. The phylogenetic relationships of the gene models were evaluated using the neighbor-joining or minimum-evolution tree with 1,000 bootstrap trials. The full-length amino acid sequences were aligned by using ClustalW33 and displayed with DNAMAN. The exon-intron structures were displayed by using the Gene Structure Display Server (GSDS, http://gsds.cbi.pku.edu.cn/index.php).

Results

In silico identification of Salix TUA and TUB genes

The present study identified a total of eight TUA genes, which were designated SaTUA1 through SaTUA8 and 20 TUB genes, namely, SaTUB1 to SaTUB20. The size of the predicted open reading frames of the eight TUA genes ranged from 1,350 bp to 1,356 bp, whereas that of the TUB genes ranged from 1,335 to 1,356 bp. The eight TUA cDNAs encode eight distinct TUA proteins, whereas the 20 TUB cDNAs encoded 19 TUB proteins, mainly because TUB7 and TUB12 encoded the same protein. The length of the TUA proteins ranged from 449 to 451 amino acids, whereas that of the TUB proteins ranged from 444 to 451 amino acids. The Salix TUAs shared 73.9% to 94.5% cDNA sequence and 88.6% to 98.4% protein sequence identity (Supplementary file 2: Figure S1, Table S2), whereas the TUBs shared 74.6% to 99.8% cDNA and 86.8% to 99.1% (except for SaTUB7/12) protein sequence identity (Supplementary file 3: Figure S2, Table S3).

Three functional domains in TUA and TUB were characterized using electron crystallography: the N-terminal domain, which contained the GTP binding site; the C-terminal domain, which comprised microtubule-associated proteins (MAPs); and the motor protein binding region and an intermediate domain containing the Taxol binding site34. PTMs are essential for the maturation of the tubulin protein, which include modifications such as tyrosination/ detyrosination35, acetylation36, polyglutamylation37, phosphorylation and polyglycylation38. Except for acetylation, all modifications take place in the hypervariable C-terminal region39. The Salix C-terminal region consisted of about 20 amino acid residues that constitute a major variable domain for TUB and to a lesser extent, for TUA as well. Specific differences in the C-terminal domain were detected among Salix and other plants and most plant TUA proteins are of the Y-type, wherein the last amino acid is a Tyr residue and is involved in PTMs of tyrosination/detyrosination. Three of the eight Salix TUA proteins are of the Y-type, whereas the other five terminates are TUA3-Met and TUA7-Met (M-type), TUA5-Leu (L-type) and TUA2-Glu and TUA4-Gln (E/Q-type). A different residue between Salix and Populus was observed and the last amino acid of TUA5 in Salix was a Leu residue, whereas in Populus, this was a Met residue (Fig. 1).

Figure 1
figure 1

Alignment of TUA and TUB C-terminal amino acid sequences of Salix and Populus.

Different colored boxes indicate various classes of TUAs. The GenBank accession numbers or locus of JGI of the Populus is provided in Supplementary file 4: Table S4.

Cloning of willow TUA and TUB genes

TUA and TUB genes were isolated from three different willow species, namely, S. arbutifolia, S. matsudana and S. matsudana var. tortuosa (corkscrew willow). No sequence differences were observed among these species (GenBank accession numbers of Salix TUAs are from KC238439 to KC238446 and the accession number of TUBs are from KC243679 to KC243698). The eight TUA genes were divided into two classes, namely, class I and class II, which differed both in the number and position of the exons. Class I TUA genes comprised four exons, whereas those of class II consisted of five exons (Fig. 2). In each class, the sizes of the exons were conserved compared to their homologs in the same position. Populus TUA genes (PtTUA1–PtTUA8) and Salix TUA genes (SaTUA1–SaTUA8) have the same number and position of exons, whereas these differ in terms of the size of introns, thereby resulting in variations in gene size. Compared to TUA class I genes, TUA class II genes formed long introns, particularly TUA2 and TUA8, which result in large-sized of TUA2 and TUA8 genes. SaTUB genes are divided into five classes; all of these have three exons and the size of each exon is the same, whereas two introns are located at a conserved position, which is analogous to the gene structure of Arabidopsis and Populus TUB gene family10,18. Class I TUB genes were bigger in size compared to class I-like TUB genes because the second intron in class I were longer than that in class I-like. SaTUB11 is much longer than other SaTUBs in class II because of the longer first intron. SaTUB19 and SaTUB20 genes were longer than their counterpart PtTUB19 and PtTUB20 genes in class III and SaTUB7 gene was longer than SaTUB8, which shows the opposite features compared to Populus in class IV.

Figure 2
figure 2

Salix TUA and TUB gene structures.

Different colors indicate various classes. (A) Classes I and II of TUAs. (B) Class I, class I-like, class II, class III and class IV of TUBs.

Phylogenetic analysis of the TUA and TUB families

The results of hylogenetic analysis of plant TUA proteins are presented in Fig. 3. Plant TUAs were divided into two classes, Class I and Class II, with the eight Salix, four rice, eight Populus and six Arabidopsis isoforms equally distributed between the two classes. Each Salix TUA has a corresponding homologue in Populus and the homologs share 89.3%–99.7% amino acid sequence identity (Supplementary file 4: Table S5), indicating that each pair of Salix and Populus TUAs originated from the same ancestor. In class I, most monocot and dicot TUAs were separately clustered. In the second branch of TUA class I, all of the TUAs are monocots, which include two Zea mays TUAs (ZmTUA1 and ZmTUA2), two Oryza sativa TUAs (OsTUA1 and OsTUA2), a Setaria viridis TUA (SvTUA2), a Eleusine indica TUA (EiTUA1) and two Hordeum vulgare TUAs (HvTUA2 and HvTUA3). In the third branch of TUA class I, all of the TUAs are dicots, which include eight TUAs from Salix and Populus, five TUAs from Gossypium hirsutum and four TUAs from Betula pendula, Pseudotsuga menziesii, Medicago truncatula and Prunus dulcis. The first branch of TUA class I is a mixed branch, which included both monocot and dicot species, indicating that the TUA genes in this branch were more original than the second and the third branch TUA genes. The xylem-originating SaTUA1 was clustered with PtTUA1, GhTUA2, GhTUA3 and GhTUA4 from cotton, which is involved in secondary cell wall development13. In class II, most of the monocot and dicot TUAs were also separately clustered into different branches such as the five monocot TUAs that were clustered in the first branch and dicot SaTUA2/4, which were more closely clustered with dicot AtTUA3/5, PtTUA2/4 and StTUA2 in the second branch.

Figure 3
figure 3

Phylogenetic analyses of plant TUA proteins.

Unrooted phylogenetic tree of TUA from S. arbutifolia (Sa), Populus trchocarpa (Pt), Arabidopsis thaliana (At), Betula pendula (Bp), Medicago truncatula (Mt), Eleusine indica (Ei), Gossypium hirsutum (Gh), Hordeum vulgare (Hv), Oryza sativa (Os), Prunus dulcis(Pd), Pseudotsuga menziesii (Pm), Setaria viridis (Sv), Solanum tuberosum (St), and Zea mays (Zm). GenBank accession numbers or locus (Populus) of the sequences are provided in Supplementary file 5: Table S6. Different colors indicate various species.

The SaTUBs and PtTUBs shared 88.4% to 100% amino acid sequence identity (Supplementary file 4: Table S5) and have similar properties of number and position of exons (Fig. 2), suggesting that the Salix TUBs and Populus TUBs were derived from the same ancestor. However, not every Salix TUB has a homolog in Populus, which suggests that Salix and Populus TUBs separately underwent a different type of expansion. The phylogenetic NJ tree is presented in Fig. 4 shows the classification of plant TUBs into five groups, namely, Class I, Class I-like, Class II, Class III and Class IV. Five Salix TUBs showed no homologs in Populus, which included TUB3, TUB4, TUB7/12, TUB8 and TUB11 and were clustered in classes II and IV. Class I and Class I-like comprise the largest woody plant cluster that mainly consist of Salix (8 out of 19) and Populus (10 out 20) TUB family members. PtTUB11 paired with PtTUB12 in the class I-like group, whereas SaTUB11 showed no homolog in Salix, which further indicates that PtTUB12 was derived from PtTUB11 via tandem/terminal fusion or translocations. On the other hand, SaTUB11 underwent a different evolutionary route compared to PtTUB11 and PtTUB12 after the salicoid genome-wide gene duplication. In class I-like, EgTUB1 was possibly involved in the determination of the orientation of cellulose microfibrils in plant secondary fiber cell walls40 and the inclusion of Salix TUB9 and TUB10 in this group suggests that these two TUBs might have evolved for secondary cell wall development. Classes III and IV each contain two Salix TUBs, together with two Populus and two Arabidopsis isoforms that were derived from genome-wide duplications.

Figure 4
figure 4

Phylogenetic analysis of plant TUB proteins.

Unrooted phylogenetic tree of TUB from S. arbutifolia (Sa), Zinnia elegans (Ze), Triticum aestivum (Ta), Populus trchocarpa (Pt), Arabidopsis thaliana (At), Betula pendula (Bp), Eucalyptus grandis (Eg), Lycopersicon esculentum (Le), Eleusine indica (Ei), Gossypium hirsutum (Gh), Hordeum vulgare (Hv), Oryza sativa (Os), Prunus dulcis(Pd), Pseudotsuga menziesii (Pm), Setaria viridis (Sv), Solanum tuberosum (St) and Zea mays (Zm). The five classes are indicated in different colors; the GenBank accession numbers of the sequences are provided in Supplementary file 5 Table S6.

Gene-specific transcript abundance of the TUA and TUB family in Salix

The gene-specific qRT-PCR primers were designed based on the 3’-UTRs of the eight TUA and twenty TUB genes and stem developing phloem, fully expanded leaf, stem developing xylem, shoot tip and inflorescence were analyzed by quantitative real-time PCR. Transcripts of all TUA members, with the exception of TUA5, were abundant in shoot. The class I TUA members are comparatively less abundant in all of the tissues compared to the class II TUA members. In class I, TUA1 showed the most abundant transcript in the developing xylem and was also expressed at moderate levels in the phloem, leaves and inflorescences. The M-type TUA3 and TUA7 were shoot- and leaf-specific, whereas the L-type TUA5 was less abundant in all tissues (Fig. 5). In class II, TUA2 was highly expressed in all of the tissues, followed by TUA8, whereas TUA4 and TUA6 were downregulated. In general, the Y-types TUA1, TUA6 and TUA8 were abundant in the xylem, suggesting that these originated as a xylem-specific cDNAs. The transcript pattern of Salix TUBs differed from that of TUAs. TUB class I, which included TUB1318 and TUB class IV, which consisted of TUB7, TUB8 and TUB12, showed high transcript levels compared to that of the other TUB genes. TUB12, TUB14, TUB15 and TUB19 represented the predominant TUB species in shoot tips, followed by TUB10 and TUB13 and TUB8 and TUB11 were most abundant in inflorescences. The expression pattern of each pair of TUB paralogs differed such as the pair TUB9 and TUB10, wherein TUB9 was not abundant in all tissues, whereas TUB10 was upregulated in the xylem and shoots. In the case of paralog pair TUB19 and TUB20, TUB19 was upregulated in all tissues compared to TUB20, which was only abundant in inflorescences. Such differences in expression pattern between Salix paralogs were suggestive of the functional divergence or redundancy of TUB genes.

Figure 5
figure 5

qRT-based tissue-specific transcript levels in various Salix tissues.

Tissues were collected from three trees and three biological replicates were prepared from each sample. Mean transcript copy numbers and standard error of the means from three technical replicates are small. (A) Transcript copy numbers of TUA genes in each nanogram of total RNA from different tissues. (B) Transcript copy numbers of TUB genes in each nanogram of total RNA from different tissues.

Discussion

Six TUA and nine TUB genes from Arabidopsis and five TUA and nineteen TUB genes from cotton expanded into eight TUA and twenty TUB genes in Populus and Salix, and together with the phylogenetic analysis of their deduced amino acid sequence, suggest that the Salix TUA and TUB genes underwent different expansions via eurosid genome-wide gene duplication and salicoid genome-wide gene duplication events, followed by reciprocal tandem/terminal fusion, translocations and other chromosomal rearrangements. The Salix and Populus TUA and TUB genes have identical features, including gene numbers, gene structure and amino acid changes at the C terminus, together with the very high levels of sequence identity, which demonstrate that the Salicaceae TUAs and TUBs were derived from the salicoid genome-wide gene duplication event, which was estimated to have occurred sometime between 6 and 10 million years ago41.

Plants are unable to tolerate large imbalances in the ratio of TUA to TUB within the cytoplasm. In maize, when a single tubulin gene was transformed and overexpressed, the plant did not regenerate, but when both TUA and TUB were transformed and the two heterodimer subunits were overexpressed, regeneration occurred42,43. In aspen, the transgenically manipulated TUA1 expression also failed to produce viable transformants18. The lethality caused by subunit imbalance seems to be a general phenomenon. In mammals and in yeast, overexpression of TUA or TUB leads to cell-cycle arrest and reduced cell viability44,45. In sum, the balance in the expression level of both TUA and TUB are essential to the tubulin 1:1 heterodimer assembly, which is crucial for the growth and development of plants and animals. In mammals, the TUA and TUB families are identical in size, seven genes of each family were identified in human and mouse46,47, thus facilitating the formation of the tubulin 1:1 heterodimer. On the other hand, in the present study involving Salix, the size of the of the TUA and TUB families differed (eight TUAs and nineteen TUBs), thus rendering a more difficult situation for the formation of the 1:1 heterodimer, because of the higher number of TUB genes compared to that of the TUAs. The mechanism responsible for the difference in size of the two families in Salix is unknown; however, there appears to be a solution to the establishment of the 1:1 heterodimer assembly. The transcript copy numbers between TUAs and TUBs are markedly different, wherein the transcript copy number of TUAs ranged from 153 to 89,271 per nanogram of total RNA, whereas that of the TUBs ranged from 4 to 16,156 per nanogram of total RNA. This observation indicates that the TUAs have multiple-fold higher number of transcripts than TUBs (Fig. 5), which apparently compensates for the lower number of members in the TUA gene family.

Cells generate distinct microtubule subtypes by two ways, namely, by expression of different tubulin isotypes and by PTMs such as detyrosination and further cleavage to Δ2-tubulin, acetylation, polyglutamylation and polyglycylation48. Except for acetylation, all of the tubulin PTMs occurs at the C terminal tails, which are exposed at the outer surface of microtubules of TUA and TUB (Fig. 6). Molecular cloning of tubulin genes showed that a C-terminal Tyr is encoded by most TUA genes49, the tyrosination–detyrosination cycle is initiated by the removal of a Tyr functional group (detyrosination) and re-addition of Tyr (tyrosination) serves to reverse the modification and to return tubulin to its initial state49,50. Salix TUAs have a distinct C-terminal residue compared to the other species and four different types of TUAs were identified in the present study, namely, three Y-types, which include M-type, L-type and E-type and the Q-type. Except for the three Y-type TUAs, the second to the last amino acid of the other TUAs are still the Y residues (Fig. 1).

Figure 6
figure 6

Putative tubulin C-terminal PTMs in Salix.

Schematic representation of the generation and removal of each tubulin PTMs with corresponding enzymes and amino acids involved in PTMs are shown in the box at the top right corner. (A) Glutamylation and polyglutamylation. (B) Glycylation and polyglycylation. (C) Putative deleucylation and demethiolation. (D) Putative deglutamynation, deglutamylation and deaspartylation.

Detyrosination is the first step of C-terminal PTM and it increases the stability and longevity of MTs. It is also essential for motors and non-motor MAPs binding to MTs51,52. In Salix, the additional amino acids are blocks of detyrosinases and the tubulin heterodimer requires other enzymes to remove these amino acid residues to achieve the detyrosinate state. We hypothesized that in Salix, other PTMs are responsible for the removal of the additional amino acid residues prior to detyrosination; therefore, the none Y-type TUAs were detyrosinated after completion of the other putative PTMs. These PTMs are termed deleucylation, demethiolation, deglutamynation and deglutamylation (Fig. 6). The putative deleucylases possibly removes the Leu residue present at the C-terminal of TUA5, demethiolation removes Met residues at the C-terminal of TUA3, TUA7 and the putative deglutamynases apparently catalyzes the removal of Gln residues of the C-terminal of TUA4. Detyrosinated tubulin can be further converted to Δ2-tubulin by removal of a Glu residue from the C-terminal, thereby exposing this region after detyrosination that is catalyzed by deglutamylase enzymes25,53,54. Δ2-tubulin generally occurs at the final stages of PTMs and Δ2-modification irreversibly locks microtubules in a detyrosinated state, thereby excluding the MTs from depolymerization-polymerization cycles55,56. Similarly, Δ2-modification has the biochemical function of regulating MT dynamics and stability. In the Salix TUA family, two distinct amino acid residues occur right before the Y residue; in class I, the residue is Glu, whereas in class, this corresponded to the Asp residue. The Glu residue can be removed by deglutamylase enzymes, thereby forming Δ2-tubulin25, whereas no enzyme catalyzed Asp to form Δ2-tubulin and the Asp showed similar chemical properties to that of Glu, wherein both are acidic and electronegative. The putative deaspartylation involves the removal of the Asp residue, thereby forming Δ2-tubulin, which is similar to that of deglutamylation as presented in Fig. 6. These additional PTMs of the TUA gene family facilitate in the formation of different tubulin heterodimers that provides functional diversity for adaptation for the woody perennial growth of Salix in its natural habitat.

Additional Information

How to cite this article: Rao, G. et al. Characterization and putative post-translational regulation of α- and β-tubulin gene families in Salix arbutifolia. Sci. Rep. 6, 19258; doi: 10.1038/srep19258 (2016).