In silico analysis of the structural diversity and interactions between invertases and invertase inhibitors from potato (Solanum tuberosum L.)

Abstract

We performed sequence diversity, phylogenetic profiling, 3D structure modelling and in silico interactions between invertases (cell wall/apoplastic and vacuolar) and invertase inhibitors (cell wall/apoplastic and vacuolar) from potato. Cloning and sequencing of invertase inhibitors was performed from different potato cultivars. The comparison of the protein sequences of the different isoforms of invertases and invertase inhibitors exhibited insertions and deletions as well as the variation in terms of amino acid residues. Furthermore, the phylogenetic tree analysis displayed two groups of invertase inhibitors corresponding to the cell wall/apoplast and vacuole. Using Phyre² protein homology recognition engine, it revealed that the structure of invertase inhibitors was predominantly α-helical and that of invertase was α helices and β strands. Results of the Ramachandran plots for each structure showed that the percentage of amino acid residues in favoured region and in allowed region. Also, the Z score and QMEAN score indicated overall good, acceptable and reliable models. In silico interactions between different isoforms of invertase and invertase inhibitors suggested that cell wall/ apoplastic invertase inhibitor exhibited stronger interaction with vacuolar invertase compared to the vacuolar invertase inhibitor. In silico interactions provides valuable information in selecting the appropriate combinations of invertase and invertase inhibitor. Therefore, a better understanding of the interactions between specific invertase and invertase inhibitor alleles will be helpful for an intelligent manipulation of the cold-induced sweetening process of potato tubers.

Electronic supplementary material

The online version of this article (10.1007/s13205-020-02171-y) contains supplementary material, which is available to authorized users.

Introduction

Plant invertases (β-fructofuranosidase, EC 3.2.1.26) convert sucrose into reducing sugars (RS) (glucose and fructose) and are essential for plant metabolism and development. Depending on their subcellular localization, higher plants contain three invertase isoenzymes viz. cell wall/apoplastic, vacuolar and cytosol (Siemens et al. 2011). The activity of invertases (cell wall/apoplastic and vacuolar) are regulated at the post-translational level via compartment-specific inhibitory proteins (cell wall/apoplastic and vacuolar), referred to as invertase inhibitors. Invertase inhibitors belong to the pectin methylesterase inhibitors (PMEIs) family with almost identical structural properties. Moreover, invertase inhibitors are ~ 17 kDa proteins, heat-stable and non-glycosylated monomers affecting invertase activity in a pH-dependent manner (Scognamiglio et al. 2003; Hothorn et al. 2004a, b).

In potato, both cell wall/apoplastic (CIF or INH1) and the vacuolar (VIF or INH2) isoforms of inhibitors play an important role in resistance to cold-induced sweetening (CIS), which causes a great loss to the potato processing industry (Brummell et al. 2011; Datir et al. 2012). Starch stored in potato tubers gets converted into RS in the sucrose hydrolysis process via the vacuolar invertase enzyme. Such high sugar potatoes produce unacceptable brown to black coloured potato chips upon frying at high temperatures (Dale and Bradshaw 2003). Therefore, the manipulation of invertase activity at the post-translational level through inhibitory proteins may help to solve the CIS problem in potato tubers. Overexpression of the INH2 gene in transgenic potato plants resulted in a 77% reduction in RS and was found to be an effective approach to control the process of CIS in potato tubers (Cheng et al. 2007). Sequencing and characterization of both INH1 and INH2 genes from cold-susceptible and resistant potato cultivars revealed that during cold storage, the transcripts of invertase inhibitors were accumulated to higher abundance in potato cultivars resistant to CIS than in susceptible cultivars (Brummell et al. 2011). Their experiment led to the conclusion that increased amounts of invertase inhibitor may contribute to the suppression of acid invertase activity and prevent the cleavage of sucrose. Interaction between invertase and invertase inhibitor proteins in cold-stored potato tubers suggested that a protein complex is underlying post-translational regulation of invertase (Lin et al. 2013). In vitro studies conducted by Sander et al. (1996) demonstrated that I_NT (apoplastic invertase inhibitor from tobacco) and I_LE (vacuolar invertase inhibitor from tomato) inhibit the cell wall/apoplastic invertase and vacuolar invertase. Their experiment showed a much faster complex formation between NtCIF (apoplastic invertase inhibitor from tobacco) and NtCWI (apoplastic invertase from tobacco), compared to NtCIF and the vacuolar invertase from tomato. Further, their investigation led to the conclusion that the mechanism of inactivation was clearly different for the cell wall/apoplastic invertase and the vacuolar invertase. Thus, the regulation of interaction between the cell wall/vacuolar invertases and INH1/INH2 and the inhibitory mechanism are of significant interest, not only in plant physiological processes, but more specifically to the potato processing industry. Invertase inhibitors have been sequenced and characterized from potato (Brummell et al. 2011; Datir et al. 2012, 2019); however, the mechanisms of how vacuolar invertase and cell wall/apoplastic invertase are inactivated by their respective inhibitors are unknown. An interaction between potato cell wall/apoplastic invertase and cell wall/apoplastic invertase inhibitor demonstrated that the invertase inhibitor is potentially an important candidate in the process of CIS (Baldwin et al. 2011). An interaction between Arabidopsis cell wall/apoplastic invertase 1 and INH1 from tobacco indicated a conserved amino acid motif PKF (Proline, Lysine and Phenylalanine) in INH1 directly targets the invertase active site (Hothorn et al. 2010). Though these invertase inhibitors are known to regulate the invertase activity and the downstream Glc–signalling cascades in vivo (Balibrea Lara et al. 2004; Jin et al. 2009), it is not fully understood how they recognize and inhibit their enzyme targets (Hothorn et al. 2010). There are no reports on in silico interaction between the cell wall/vacuolar invertase by their inhibitors. The crystal structure of the complex between INV1 and the cell wall/apoplastic invertase inhibitor INH1 from tobacco reveals that INH1 directly inserts a small amino acid motif into the invertase substrate binding cleft. Hence, it can be expected that the INH2 might use similar surface areas and essentially the same targeting mechanism to inhibit vacuolar isoenzymes (Hothorn et al. 2010). Due to the limited availability of pure protein components, there has been a major obstacle in the biochemical and structural investigation of invertase–inhibitor interaction. Therefore, the resolution of the invertase–inhibitor structure is another step towards a better understanding of the regulation of invertase–inhibitor complex (Hothorn et al. 2004b).

Screening and identification of potato cultivars resistant to CIS requires long-term storage of potato tubers at cold temperatures. Identification of the CIS resistance potato genotypes is mainly determined by the amount of RS and chip/crisp colour of long-term cold-stored potato tubers. However, this procedure is tedious and time consuming. Identification and selection of CIS resistant genotypes can be performed at the molecular level by cloning, sequencing and identifying SNP diversity for invertase and invertase inhibitor genes from diverse potato germplasm. Once SNP diversity is identified, variations in the protein sequences of invertases and invertase inhibitors can be predicted. Then in silico interactions between invertases and invertase inhibitors can be performed using bioinformatics tools, and the proper combinations of invertase and invertase inhibitors can be determined. It can be presumed that the microheterogeneity, i.e. variation in the amino acid sequence of the protein arising due to the SNPs in the cell wall/ apoplastic and vacuolar invertase inhibitor genes could be associated with the variation in the RS levels of CIS resistant and susceptible potato cultivars (Datir et al. 2012, 2019). Therefore, considering the importance of invertases and invertase inhibitors in the CIS process of potato, the present investigation was focused on the bioinformatics analysis of the structural diversity, identification of conserved motifs and phylogenetic relationship analysis of invertases and invertase inhibitors from potato. To date, three-dimensional structures of invertase and invertase inhibitor proteins are not available. Therefore, we have performed homology modelling studies for these proteins by using the sequences from potato. Homology modelling approach is widely used to predict the tertiary structure of proteins as well as the structures of interacting proteins and it is based on the principle that homologous proteins (proteins sharing similar sequences) should have similar structures (Chothia and Lesk 1986). To predict the 3D structures of invertase and invertase inhibitor of potato, a crystal structure of cell wall invertase from Arabidopsis thaliana in complex with a specific invertase inhibitor from Nicotiana tabacum was used as a template structure. Predicted interactions between different isoforms of invertases (cell wall/apoplastic and vacuolar) with that of different isoforms of invertase inhibitors (cell wall/apoplastic and vacuolar) were studied based on the non-bonded interactions in the complexes in order to comprehend how invertases coordinate responses to both cell wall/apoplastic and vacuolar inhibitor isoforms. A better understanding of the interactions between these two proteins will facilitate the intelligent manipulation and selection of appropriate combinations of invertase and invertase inhibitors involved in the CIS process of potato tubers.

Materials and methods

The flow chart for materials and method used in computational steps to reach protein–protein interactions is depicted in Supplementary Fig. S1.

Cloning and sequencing of the invertase inhibitors from potato

INH1 (GenBank accessions AY864819, AY864820, GU980592) and INH2 (GenBank accession XM_006349784) genes specific primers were designed based on potato sequences available in the GeneBank according to Rozen and Skaletsky (2000) using Primer 3 software (https://biotools.umassmed.edu/bioapps/primer3_www.cgi). Products were amplified from genomic DNA extracted from the leaf of different potato cultivars (Datir et al. 2012, 2019) using the INH1 primers INH1F1 (5′-CACATTTAGTTCTTAATTTTCCCAA-3′) and INH1R1 (5′-AGAAGGGGACAAACATATATTGGA-3′) and the INH2 primers INH2F1 (5′-CCTTCATCAACTTCTCATTTCTTC-3′) and INH2R1 (5′-GTGCATTGAACGGCAAATTA-3′). The primers were based on 5′ and 3′ UTRs of the gene. Further, PCR isolation of fragments, cloning and sequencing of INH1 and INH2 were described in detail by Datir et al. (2012, 2019). Both INH1 and INH2 sequences were submitted to NCBI under gene bank accession numbers AFI47459 and AYV96512 respectively.

Retrieval of invertase and invertase inhibitor sequences

In the present study, the amino acid sequences of the cell wall/apoplastic invertase (CWI), vacuolar invertase (VI), cell wall/apoplastic invertase inhibitor (INH1) and the vacuolar invertase inhibitor (INH2) from potato were used for the bioinformatics analysis. Five amino acid sequences of the cell wall/apoplastic invertase (CWI-Like—XP_006342970, Inv_ap-a pCD141—AEV46339, Inv_ap-a pCD111—PGSC0003DMT400006639/ADN06440, Inv_ap-b InvGE—AEV46300, Inv_ap-b InvGF—AEV46318), one VI—vacuolar invertase Pain 1 (ADM47340), one cell wall/apoplastic invertase inhibitor—INH1 (AFI47459) and one vacuolar invertase inhibitor—INH2 (AYV96512) from Solanum tuberosum were retrieved from the sequence database of NCBI (https://www.ncbi.nlm.nih.gov) and The Potato Genome Sequencing Consortium (The Potato Genome Sequencing Consortium 2011) (Table 1). Orthologous sequences that were hypothetical proteins at the time of work and corresponded to the proteins in potato were obtained through The Potato Genome Sequencing Consortium (2011), BLASTp and Sci-BLAST programs (Altschul et al. 1990). The information along with their gene bank accession number, linkage group and length of the amino acids of selected invertase and invertase inhibitor sequences is presented in Table 1.

Table 1.

Information on selected invertases and invertase inhibitors

Sr. no	Gene name	Loci	Chromosome	Accession Number	Amino acid
1	Beta-fructofuranosidase (CWI) cell wall invertase	CWINV1-like	III	XP_006342970	571
2	Cell wall/apoplastic invertase (CWI)	Invap-a pCD141	X	AEV46339	582
3	Cell wall/apoplastic invertase (CWI)	Invap-a pCD111	X	ADN06440	590
4	Cell wall/apoplastic invertase (CWI)	Invap-b InvGE	IX	AEV46300	586
5	Cell wall/apoplastic invertase (CWI)	Invap-b InvGF	IX	AEV46318	581
6	Vacuolar invertase (VI)	Pain-1	III	ADM47340	639
7	Cell wall/apoplastic invertase inhibitor (INH1/CIF)	INH1	XII	AFI47459	171
8	Vacuolar invertase inhibitor (INH2/VIF)	INH2	XII	AYV96512	181

Invertase and invertase inhibitors from potato used in the present study along with their gene bank accession number, linkage group and length of the amino acids are represented

Alignment and comparison of invertase and invertase inhibitors sequences

Alignment and comparison of the selected protein sequences of invertase and invertase inhibitors was performed in MEGA 6.0 using ClustalW (https://www.genome.jp/tools-bin/clustalw (Thompson et al. 1994) using default parameters. All the protein sequences are obtained from the GenBank database (https://www.ncbi.nlm.nih.gov/). Invertase inhibitor sequences used in the alignment are; StVIF—Solanum tuberosum vacuolar invertase inhibitor—AYV96512, StCIF—Solanum tuberosum cell wall/apoplastic invertase inhibitor—AFI47459, SlVIF—Solanum lycopersicum vacuolar invertase inhibitor—NP_001316149, NtVIF—Nicotiana tabacum vacuolar invertase inhibitor—AAN60076, IbVIF/CIF—Ipomea batata vacuolar/ cell wall invertase inhibitor—AAM94391, SlCIF/VIF—Solanum lycopersicum cell wall invertase inhibitor—NP_001234791, NtCIF—Nicotiana tabacum cell wall invertase inhibitor—CAA73333, Br CIF/VIF—Brassica rapa cell wall/vacuolar invertase inhibitor—XP_009148010, BnCIF/VIF—Brassica napus cell wall/svacuolar invertase inhibitor—XP_013641163, Bo CIF/VIF—Brassica oleracea cell wall/vacuolar invertase inhibitor—XP_013589892, AtVIF—Arabidopsis thaliana vacuolar invertase inhibitor—NP_001320618, GhCIF/VIF—Gossypium hirsutum cell wall/vacuolar invertase inhibitor—XP_016676243 and TcCIF/VIF—Theobroma cacao cell wall/vacuolar invertase inhibitor—XP_017974170. The secondary structure of the invertase inhibitors was predicted using PSIPRED v3.3 and compared with that of previously published information from N. tabacum—Nt-INH1/ NtVIF (Greiner et al. 1998, 1999; Hothorn et al. 2004a, 2010; Reca et al. 2008).

The alignment and comparison of selected invertase sequences from potato was performed to identify conserved motifs and deletions. The sequences included are; cell wall/apoplastic invertase (CWI-Like—XP_006342970, Inv_ap-a pCD141—AEV46339, Inv_ap-a pCD111—PGSC0003DMT400006639/ ADN06440, Inv_ap-b InvGE—AEV46300, Inv_ap-b InvGF—AEV46318) and VI—vacuolar invertase Pain 1 (ADM47340).

Construction of the phylogenetic tree from invertase inhibitor sequences from planta

A phylogenetic tree was constructed using the cell wall/apoplastic invertase inhibitor and the vacuolar invertase inhibitor sequences using MEGA (Molecular Evolutionary Genetic Analysis) 6.0 programme (Tamura et al. 2011) by neighbour-joining method using 1000 bootstrap iterations. The protein sequences used in the construction of phylogenetic tree are; the cell wall/apoplastic invertase inhibitor—INH1 (StCIF: AFI47459) and the vacuolar invertase inhibitor—INH2 (StVIF: AYV96512) from Solanum tuberosum, Solanum lycopersicum (SlVIF: NP_001316149 and SlCIF: NP_001234791), Nicotiana tabaccum (NtVIF: XP_016474343 and NtCIF: CAA73333), Ipomea batata (IbCIF/VIF: AAM94391), Brassica rapa (BrCIF/VIF: XP_009148010), Brassica napus (BnCIF/VIF: XP_013641163), Brassica oleracea (BoCIF/VIF: XP_013589892), Arabidopsis thaliana (AtVIF1: NP_001320618), Gossypium hirsutum (GhCIF/VIF: XP_016676243) and Theobroma cacao (TcCIF/VIF: XP_017974170). All the protein sequences are obtained from the GenBank database (https://www.ncbi.nlm.nih.gov/).

Homology modelling of protein structures of invertase and invertase inhibitors from potato

To identify the possible structural homologs, sequences of invertase and invertase inhibitors were subjected to protein BLAST against the protein data bank (PDB). A crystal structure with a resolution of 2.58 Å of plant cell wall invertase (A. thaliana) [PDB ID: 2XQR, Chain A] in complex with a specific invertase inhibitor (Nicotiana tabacum) [PDB ID: 2XQR, Chain B] was obtained from protein data bank (https://www.rcsb.org/) and selected as the template for building models. Homology models of invertase and invertase inhibitors from Solanum tuberosum L. were built using the above template structures (2XQR_A for invertase proteins and 2XQR_B for invertase inhibitors) with the help of the Swiss-Model webserver (https://www.expasy.org/swissmod/SWISS-MODEL.html). The compositions of the protein structures of invertases and invertase inhibitors were further built using Phyre² (Protein Homology/analogy Recognition Engine V 2.0—Tyagi et al. 2012) and compared with the template and visualized using PyMOL. Individual protein sequences of invertases (cell wall/apoplastic invertase—CWI-Like—XP_006342970, Inv_ap-a pCD141—AEV46339, Inv_ap-a pCD111—PGSC0003DMT400006639/ ADN06440, Inv_ap-b InvGE—AEV46300, Inv_ap-b InvGF—AEV46318 and VI—vacuolar invertase Pain 1 -ADM47340) and invertase inhibitors (GeneBank Accession numbers INH1-AFI47459 and INH2-AYV96512) were submitted to Phyre2 search engine under intensive mode (https://www.sbg.bio.ic.ac.uk/phyre2/). The predicted models compared for helices, beta sheets and loops. The stereo chemical quality of each model was evaluated with PROCHECK (Laskowski 1993) and the Ramachandran plot (Ramachandran et al. 1963). The Ramachandran plot is a plot of the torsional angles—phi (φ) and psi (ψ)—of the residues (amino acids) contained in a peptide. Making a Ramachandran plot helps to determine which torsional angles are permitted and can obtain insight into the structure of peptides (Ramachandran et al. 1963). The model with satisfying stereo-chemical quality was further assessed by QMEAN (Benkert et al. 2009) and ProSA (Wiederstein and Sippl 2007). ProSA was used to compute Z score values for the models.

The modelled structures were assessed by several structure validation programs like PROCHECK, ProSA and Qmean. Ramachandran plots were obtained from PROCHECK. Ramachandran plot can check the overall stereochemical quality of a protein structure by plotting phi(Φ)–psi(ψ) torsion angles for every residue of a protein. Modelled structures used in this study were validated by PROCHECK program and Ramachandran maps were generated using Discovery studio (Dassault Systemes BIOVIA 2016). The details of PROCHECK analysis in terms of residues present in most favoured regions, in allowed regions, in generously allowed regions and in disallowed regions are tabulated in Table 4. The QMEAN Z score measures the absolute quality of a model by comparing its QMEAN score to scores of experimentally solved proteins (Benkert et al. 2009). QMEAN scores of final models are also provided in Table 4 and all scores fall within the acceptable range of scores found by reference structures of similar size. ProSA web algorithm performs similar analysis like QMEAN. ProSA and QMEAN are accessed through the webservers (https://prosa.services.came.sbg.ac.at/prosa.php and https://swissmodel.expasy.org/qmean/) and these two tools can also be found in expasy.org.

Table 4.

Information on Ramachandran plot for invertases and invertase inhibitors

Name	Number of residues in favoured region (%)	Number of residues in allowed region (%)	Number of residues in outlier region (%)	Z score	QMEAN
CWINV1-like	92.1	6.6	1.3	− 7.85	− 3.23
VI—Pain-1	94.4	4.8	0.7	− 8.21	− 1.28
Invap-a pCD141	92.9	6.2	0.9	− 8.53	− 2.14
Invap-a pCD111	92.5	6.0	1.5	− 8.34	− 2.06
Invap-b InvGE	93.6	4.9	1.5	− 8.17	− 1.93
Invap-b InvGF	93.4	5.5	1.1	− 8.64	− 2.09
INH1	93.2	6.1	0.7	− 6.13	− 1.83
INH2	92.0	7.3	0.7	− 8.16	− 1.95

Z score and QMEAN values are presented

Interaction prediction of invertase and invertase inhibitors

Structurally conserved regions (SCRs) between models of invertase and invertase inhibitors of Solanum tuberosum and homologous proteins (PDB: 2XQR) were aligned by PyMol (www.pymol.org) (DeLano 1997) and root mean square deviation (RMSD) representing average distance between the backbones of superimposed proteins and were checked for each protein. PyMOL is a molecular visualization program used to represent protein structures and visualize protein–protein or protein-small molecule interactions. Root Mean Square Deviation (RMSD) is the most commonly used quantitative measure of the similarity between two superimposed atomic coordinates of protein structures. Lesser the RMSD value represents similar protein structures. Here we have computed RMSD of modelled structure with respect to the template structure.

Interacting residues in the interfaces between invertase and invertase inhibitor (all possible combinations as shown in Tables 5 and 6) are identified using the PISA (Protein Interfaces, Surfaces and Assemblies server) (Krissinel and Henrick 2007). PISA is an automated method for detecting macromolecular assemblies based on the analysis of interfaces and stability of assemblies reported in crystal structures (Tyagi et al. 2012).

Table 5.

Interactions of hydrogen bonds: cell wall invertase inhibitor (INH1) with cell wall invertase (CWI) and vacuolar invertase (VI)

INH1_CWIInvGE		INH1_CWIInvGF		INH1_CWILike		INH1_CWIpCD111		INH1_CWIpCD141		INH1_VI
LYS 115	TYR 123	LYS 115	PHE 121	LYS 115	LYS 111	LYS 115	TYR 122	LYS 115	TYR 122	THR 110	TYR 203
LYS 119	GLU 246	LYS 119	GLU 245	LYS 119	GLU 235	LYS 119	GLU 245	LYS 119	GLU 246	GLU 111	ASN 177
ARG 168	VAL 281	ARG 168	LEU 277	GLU 111	ASN 81	LYS 119	GLU 245	LYS 119	GLU 246	GLU 111	ASN 147
GLU 111	ASN 93	ARG 168	LEU 277	GLU 111	ASN 81	ARG 168	VAL 279	ARG 168	VAL 280	GLU 111	ASN 147
GLU 111	ASN 93	GLU 111	ASN 91	LYS 115	GLN 139	SER 97	ASN 344	GLU 111	ASN 92	GLY 116	GLN 205
LYS 115	GLN 152	GLU 111	ASN 91	GLY 116	GLN 139	GLU 111	ASN 92	GLU 111	ASN 92	LYS 119	LYS 336
GLY 116	GLN 152	LEU 113	GLN 148	GLY 116	ARG 180	GLU 111	ASN 92	LYS 115	GLN 151	ASP 123	LYS 336
GLY 116	ARG 191	LYS 115	GLN 150	LEU 171	GLN 200	LYS 115	GLN 151	GLY 116	GLN 151	LEU 171	LYS 265
LYS 119	ARG 283	GLY 116	GLN 150	LEU 171	MET 233	GLY 116	GLN 151	GLY 116	ARG 190	LEU 171	MET 298
GLU 122	ARG 283	GLY 116	ARG 189			GLY 116	ARG 190	LYS 119	ARG 282	SER 105	GLU 398
GLU 134	ASN 500	LYS 119	ARG 279			LYS 119	ARG 281	GLU 122	ARG 282	SER 105	GLU 398
LEU 171	ASN 244	GLU 122	ARG 279			GLU 122	ARG 281	LEU 170	GLN 188	LYS 115	ILE 176
LEU 171	VAL 281	LEU 171	ASN 243			LEU 170	GLN 188	LEU 171	ASN 244	LYS 119	GLU 300
		LEU 171	LEU 277			LEU 171	ASN 243	LEU 171	VAL 280	LYS 119	GLU 300
						LEU 171	VAL 279			ARG 168	ASP 334
										ARG 168	ASP 334

INH1—cell wall invertase inhibitor, CWIInvGF, CWILike, CWIpCD111, CWIpCD141—cell wall invertases, VI—vacuolar invertase. Residues THR110, Asp123 and SER105 from INH1 are uniquely interacting with VI are denoted in italics

Table 6.

Interactions of Hydrogen Bonds: vacuolar invertase inhibitor (INH2) with cell wall invertase (CWI) and vacuolar invertase (VI)

INH2_ CWILike		INH2_CWIpCD141		INH2_CWIpCD111		INH2_CWIInvGE		INH2_CWIInvGF		INH2_VI
GLY 126	GLN 139	GLY 126	GLN 151	GLY 126	GLN 151	GLY 126	GLN 152	ASP 62	LYS 211	GLU 121	ASN 177
GLU 141	GLU 333	LYS 129	ARG 282	LYS 129	ARG 281	LYS 129	ARG 283	LYS 124	GLN 148	LYS 125	TRP 180
LEU 181	GLN 200	ASP 137	TYR 501	LEU 280	GLN 188	ASP 137	TYR 504	GLY 126	GLN 150	GLY 126	GLN 205
LEU 181	MET 233	LEU 180	GLN 188	LEU 181	ASN 243	LEU 181	ASN 244	LYS 129	ARG 279	GLY 126	ARG 244
		LEU 181	ASN 244					LEU 181	ASN 243	LYS 129	LYS 336
										GLU 141	SER 399
										LEU 180	LYS 265
										LEU 181	LYS 265
										LEU 181	MET 298
										ASP 115	GLU 398
										GLY 126	ASP 199

INH2—vacuolar invertase inhibitor, CWIInvGF, CWILike, CWIpCD111, CWIpCD141—cell wall invertases, VI— vacuolar invertase

Results

Cloning and sequencing of invertase inhibitors from potato cultivars

PCR generated multiple bands for the INH1 gene while a single band was noticed for the INH2 gene from different potato cultivars based on INH1 and INH2 specific primers (Datir et al. 2012, 2019). Upon sequencing, a total of five and four alleles were found to exist for INH1 and INH2 genes respectively (Datir et al. 2012, 2019). All the sequences were deposited to NCBI databank, and two invertase inhibitor sequences (INH1-AFI47459 i.e. cell wall/apoplastic invertase inhibitor and INH2-AYV96512 i.e. vacuolar invertase inhibitor) were further explored in the present study.

Structural analysis and comparison of invertase inhibitors from Planta

The structural analysis, alignment and comparison of the potato INH1 and INH2 protein sequences with invertase inhibitors from planta performed using the CLUSTALW program exhibited 42% sequence identity (Fig. 1). The size of the INH1 and INH2 protein sequences ranged from 162 to 192 amino acid residues and encode a variable molecular mass (17.6–20.6 kDa) with pI ranging from 4.47 to 8.88 (Fig. 1; Table 2). Secondary structure prediction of the INH1 and INH2 proteins (based on Nicotiana tabaccum–cell wall invertase inhibitor: Nt-INH1/Nt-CIF) revealed a total of seven α helices. A four-helix bundle (α4–α7) preceded by an unusual N-terminal extension forming a helical hairpin along with an additional small helix was also observed (Fig. 1). The positions of two cysteine residues (grey shaded Fig. 1) in the α1 and α2 helices as well as the other two cysteine resides linking helix α5 and α6 were well conserved in all the invertase inhibitor sequences (Fig. 1). Importantly, the amino acid motif PKF (proline, lysine and pheylalanine—orange highlight and boxed) in α6 helix is highly conserved in all the invertase inhibitor sequences (Fig. 1).

Fig. 1.

Alignment of the cell wall/apoplastic invertase inhibitors and the vacuolar invertase inhibitors from planta. Amino acid sequences of the cell wall and vacuolar invertase inhibitors from potato were aligned with that of previously published sequences from planta using multiple sequence alignment in MEGA 6.0 using ClustalW (https://www.genome.jp/tools-bin/clustalw (Thompson et al. 1994) using default parameters. The amino acids are represented by a single letter code. The protein sequences listed in the diagram are; StVIF—Solanum tuberosum vacuolar invertase inhibitor (AYV96512), StCIF—Solanum tuberosum cell wall/apoplastic invertase inhibitor (AFI47459), SlVIF—Solanum lycopersicum vacuolar invertase inhibitor (NP_001316149), NtVIF—Nicotiana tabacum vacuolar invertase inhibitor (AAN60076), IbVIF/CIF—Ipomea batata vacuolar/ cell wall invertase inhibitor (AAM94391), SlCIF/VIF—Solanum lycopersicum cell wall invertase inhibitor (NP_001234791), NtCIF—Nicotiana tabacum cell wall invertase inhibitor (CAA73333), Br CIF/VIF—Brassica rapa cell wall/vacuolar invertase inhibitor (XP_009148010), BnCIF/VIF—Brassica napus cell wall/vacuolar invertase inhibitor (XP_013641163), Bo CIF/VIF—Brassica oleracea cell wall/vacuolar invertase inhibitor (XP_013589892), AtVIF—Arabidopsis thaliana vacuolar invertase inhibitor (NP_001320618), GhCIF/VIF—Gossypium hirsutum cell wall/vacuolar invertase inhibitor (XP_016676243) and TcCIF/VIF—Theobroma cacao cell wall/vacuolar invertase inhibitor (XP_017974170). The secondary structure was predicted using PSIPRED v3.3 and compared with that of previously published information from N. tabacum—Nt-INH1/ NtVIF (Greiner et al. 1998, 1999; Hothorn et al. 2004a, 2010; Reca et al. 2008). The symbols used for the secondary structure are α1-α7 helix. The invertase inhibitor peptide encodes 162 to 192 amino acids. Asterisks at the bottom of the alignment indicate identical residues. Amino acid numbers are listed on right of the alignment. Numbers 1 and 2 at the bottom of the alignment (C residues shaded in grey) denote disulfide bridges connecting the conserved two pairs of cysteine residues. Highly conserved PKF (Proline, Glycine and Phenylalanine shaded in orange) motif in the α6 helix in the known invertase inhibitors (Hothorn et al. 2010) is boxed

Table 2.

Comparison of predicted amino acid compositions, molecular weight and pI of invertase inhibitors from planta

Sr. no	Protein name	Accession number	Number of amino acids	Molecular weight (kDa)	Predicted pI
1	StVIF	AYV96512	175	20.5	4.88
2	StCIF	AFI47459	171	18.8	8.29
3	SlVIF	NP_001316149	175	19.9	5.32
4	SlCIF	NP_001234791	171	18.8	8.30
5	NtVIF	XP_016474343	172	19.1	5.08
6	NtCIF	CAA73333	166	17.8	8.67
7	IbCIF/VIF	AAM94391	192	20.6	4.47
8	BrCIF/VIF	XP_009148010	162	17.6	7.39
9	BnCIF/VIF	XP_013641163	162	17.6	7.39
10	BoCIF/VIF	XP_013589892	163	17.8	8.63
11	AtVIF1	NP_001320618	166	18.0	8.68
12	GhCIF/VIF	XP_016676243	181	19.3	8.78
13	TcCIF/VIF	XP_017974170	181	19.3	8.62

Phylogenetic analysis

A phylogenetic tree (Fig. 2) constructed in the MEGA 7 program by neighbour-joining 1000 bootstrap analysis using INH1 and INH2 protein sequences from planta revealed that the tree clearly divides into two major branches corresponding to two classes of invertase inhibitors, viz. INH1 and INH2. Group A consisted of INH2, i.e. vacuolar invertase inhibitor sequences while group B comprised INH1, i.e. cell wall/apoplastic invertase inhibitor (Fig. 2).

Fig. 2.

Phylogenetic tree of the cell wall/apoplastic invertase inhibitors and the vacuolar invertase inhibitors from planta. Phylogenetic tree was constructed in MEGA 7 program (Tamura et al. 2011) by neighbour-joining 1000 bootstrap analysis based on the cell wall/apoplastic invertase inhibitor and the vacuolar invertase inhibitor protein sequences from planta. Scale bar 0.20 marks 0.2 amino acid substitution per site. Bootstrap values (%) are indicated at each branch point. Two groups identified. Group A consists of the vacuolar invertase inhibitors from Solanum tuberosum (StVIF—AYV96512), Solanum lycopersicum (SlVIF-NP_001316149), N. tabacum (NtVIF-XP_016474343), Ipomea batata (IbCIF/VIF- AAM94391), Theobroma cacao (TcCIF/VIF-XP_017974170), Gossypium hirsutem (GhCIF/VIF- XP_016676243), Arabidopsis thaliana (AtVIF1-NP_001320618), Brassica oleracea (BoCIF/VIF- XP_013589892), Brassica napus (BnCIF/VIF- XP_013641163) and Brassica rapa (BrCIF/VIF- XP_009148010). Group B consists of the cell wall/apoplastic invertase inhibitors from Solanum tuberosum (StCIF- AFI47459), Solanum lycopersicum (SlCIF- NP_001234791) and Nicotiana tabacum (NtCIF-CAA73333). StKunitz is set as an outgroup which is Solanum tuberosum Kunitz type invertase inhibitor. The protein sequences listed in the diagram are obtained from the GenBank database (https://www.ncbi.nlm.nih.gov/)

Structural analysis and comparison of invertases

Structural comparison between protein sequences of the cell wall/apoplastic invertase with that of vacuolar invertase (Table 1) performed using CLUSTALW program revealed that the vacuolar invertase (Pain 1) was the longest sequence (639 amino acid residues) (Fig. 3). The cell wall/apoplastic invertase was encoded by 571 to 590 amino acid residues (Fig. 3). A polymorphism analysis of different isoforms of invertase exhibited several insertions and deletions (Fig. 3). The β-fructosidase motif RDP, (NDPN) and the catalytic domain (WECV/PD) of the invertase enzymes are well conserved. Furthermore, the amino acid residues interacting with that of INH1 and INH2 (highlighted in red) were then utilized in the interaction studies (Fig. 3).

Fig. 3.

Amino acid residues of invertases (cell wall/apoplastic and vacuolar) interacting with invertase inibitors. Amino acid sequence alignment of the invertases (cell wall/apoplastic and vacuolar) from potato was conducted using multiple sequence alignment in MEGA 6.0 using ClustalW (https://www.genome.jp/tools-bin/clustalw (Thompson et al. 1994) using default parameters. The amino acids are represented by a single letter code. The amino acid residues interacting with invertase inhibitors are highlighted in red. The sequences included in the alignment are: CWINV1-like—beta-fructofuranosidase (XP_006342970), Pain-1—vacuolar invertase (ADM47340), Inv_ap-a pCD141 (AEV46339), Inv_ap-a pCD111 (ADN06440), Inv_ap-b InvGE (AEV46300) and Inv_ap-b InvGF (AEV46318) are cell wall/apoplastic invertases from potato. The protein sequences listed in the diagram are obtained from the GenBank database (https://www.ncbi.nlm.nih.gov/). Asterisks at the bottom of the alignment indicate the identical residues and – indicates deletion. The acid invertases are glycoproteins and have three conserved sequence motifs: β-fructofuranosidase motif (NDPNG(A)), RDP and WECP(V)D motifs are red boxed (Lammens et al. 2008; Roitsch & Gonzalez 2004. Amino acid numbers are listed on right of the alignment

Homology modelling of protein structures

Homology models of the cell wall/apoplastic invertase, vacuolar invertase and INH1 and INH2 from Solanum tuberosum L. (Table 1) were built using 2XQR_A (for invertase proteins) and 2XQR_B (for invertase inhibitors) using the Swiss-Model webserver. Three-dimensional structures of the above-mentioned proteins were generated by a homology modelling method based on the known crystal structure (PDB ID: 2XQR). The protein structures of invertases and invertase inhibitors built using Phyre² and their comparison with the respective templates (Table 3) revealed that all the structures were modelled at > 90% accuracy. Invertases displayed 55 to 66% sequence identity with the templates, whereas; INH1 showed 89% sequence identity, INH2 exhibited 47% sequence identity with the template (Table 3). The predicted model of invertase inhibitors was predominantly α-helical (> 75%), while that of invertases were mainly composed of α helices and β strands (Table 3).

Table 3.

Results of protein modelling using Phyre²

Name	Template	Sequence identity (%)	Residue	Confidence	Secondary structure
					Disordered (%)	α-helix (%)	β-strand (%)
CWINV1-like	c2ac1A_	55	529	100	11	06	04
VI—Pain-1	c3ugfB_	66	527	100	15	06	45
Invap-a pCD141	c2ac1A_	59	526	100	09	06	48
Invap-a pCD111	c2ac1A_	59	528	100	09	06	50
Invap-b InvGE	c2ac1A_	56	530	100	10	06	49
Invap-b InvGF	c2ac1A_	55	528	100	10	06	50
INH1	d2cj4a1	89	147	100	19	81	0
INH2	d2cj4a1	47	144	100	26	77	0

The composition of the protein structures of invertase and invertase inhibitors built using Phyre² and compared with the template

The protein structure was built by Phyre2—Protein Homology/analogy Recognition Engine V 2.0 (Tyagi et al. 2012)

Furthermore, the Ramchandran plot created for each model showed percentages of residues in favoured region and in allowed region along with Z score and QMEAN score (Table 4; Supplementary Fig. S2). The graphical output of QMEAN score, Z score and local quality estimate of CWI-Like invertase is presented in Supplementary Fig S3. All the predicted models were comprised maximum number of residues (more than 90%) in the favoured region, indicating overall acceptable models (Table 4; Supplementary Fig. S2). The QMEAN score ranged from − 1.28 to – 3.23 and lies in the range of estimated model reliability values, which are between 0 and 1 (Table 4). Among all the isoforms of invertase, the maximum QMEAN score was noticed in the CWINV1-Like i.e. − 3.23 while it was lowest in the VI-Pain − 1, i.e. − 1.28 (Table 4). Z score was slightly different in the case of invertase isoforms. However, INH1 displayed − 6.13 and INH2 showed higher Z score value, i.e. − 8.16 (Table 4). The QMEAN and Z score measures the absolute quality of a model by comparing its QMEAN score to scores of experimentally solved proteins. Lesser Z score represents well fitted model for the predicted structure to the range that is typical for proteins of similar size in the protein database. ProSA also represents the quality in the energy plots and minimum values in the plot account for the nativity and stability of the molecules.

Interaction between invertases and invertase inhibitors

Results of the in silico interactions of INH1 and INH2 with different isoforms of invertase showed residues that contribute to the hydrogen bond formation in respective interacting molecules (Tables 5, 6; Figs. 3, 4; Supplementary Fig. S4). Interestingly, the INH1 was found to interact with both the isoforms of invertase, i.e. cell wall/apoplastic as well as vacuolar. Among different isoforms of the cell wall/apoplastic invertase, CWIpCD111 interact with INH1 through the hydrogen bond formation with residues like ASN92, TYR122, GLN151, GLN188, ARG190, ASN243, GLU245, VAL279, ARG281 and ASN344 (Tables 5, 6). On the other hand, nine residues (mostly homologs of the above-mentioned residues) participate in the hydrogen bond formation in CWIInvGF (Table 5). Although the INH1 was interacting with all the isoforms of cell wall/apoplastic invertase, interestingly it displayed nine residues interacting with that of vacuolar invertase (Table 5). Residues THR110, ASP123 and SER105 from INH1 were specifically found to interact with that of the vacuolar invertase and did not show interaction through the hydrogen bond formation with any of the cell wall/apoplastic invertase isoforms (Table 5). On the other hand, strong interactions were noticed between INH2 and the vacuolar invertase (majorly driven by 8 h-bond formation), whereas interactions between INH2 and cell wall/apoplastic invertase isoforms were varied and directed by weak hydrogen bond formations (Table 6).

Fig. 4.

Amino acid residues of invertase inhibitors (cell wall/apoplastic and vacuolar) interacting with invertases. Alignment of INH1—cell wall/apoplastic invertase inhibitor and INH2—vacuolar invertase inhibitor from potato was performed using multiple sequence alignment in MEGA 6.0 using ClustalW (https://www.genome.jp/tools-bin/clustalw (Thompson et al. 1994) using default parameters. The secondary structure was predicted using PSIPRED v3.3 and compared with that of previously published information from N. tabacum—Nt-INH1/ NtVIF (Greiner et al. 1998, 1999; Hothorn et al. 2004a, 2010; Reca et al. 2008). The amino acids are represented by a single letter code. Amino acid residues which are interacting with invertases are highlighted in red. The symbols used for the secondary structure are α1-α7 helix. Four conserved cysteine residues are highlighted in yellow and the numbers 1 and 2 at the bottom of the alignment denote disulfide bridges connecting the conserved two pairs of cysteine residues. Asterisks at the bottom indicate identical residues. Amino acid numbers are listed on right of the alignment. Highly conserved PKF (Proline, Glycine and Phenylalanine shaded in orange) motif in the α6 helix in the known invertase inhibitors (Hothorn et al. 2010) is boxed. The protein sequences listed in the diagram are obtained from the GenBank database (https://www.ncbi.nlm.nih.gov/)

Discussion

We have studied the structural diversity, 3D modelling and in silico interactions between different isoforms of invertases and invertase inhibitors from potato using bioinformatics tools. The results presented in this study on the in silico interactions between invertases and invertase inhibitors suggested that both the cell wall/apoplastic invertase inhibitor as well as vacuolar invertase inhibitor might be required to inhibit the vacuolar invertase in potato tubers. However, for this a proper combination of invertase and inhibitory proteins is essential. Here is the first time that some of the residues from invertase inhibitors examined in silico interacted with that of invertases. Based on these studies, SNP diversity for invertase and invertase inhibitors can be studied, the in silico interactions can be performed and the proper combinations for these genes can be identified from the diverse CIS potato germplasm.

Structural diversity and features of invertase and invertase inhibitory proteins

Structural features, comparison of invertase inhibitor protein sequences (cell wall/apoplastic invertase inhibitor and vacuolar invertase inhibitor) from planta (Fig. 1; Table 2) and the prediction of secondary structure (Nicotiana tabaccum-cell wall/apoplastic invertase inhibitor) (Greiner et al. 1998, 1999; Hothorn et al. 2003, 2004a), are in agreement with the previously published reports (Liu et al. 2010; Datir et al. 2012, 2019). The conserved cysteine residues (grey shaded Fig. 1) in α1 and α2 helices and α5 and α6 are the characteristic of plant invertase inhibitors. These cysteine residues are known to be specifically involved in the disulfide bridge formation (Hothorn et al. 2004a; Rausch and Greiner 2004). Detailed studies on the invertase inhibitor revealed that these disulfide bridges connecting to α helices in the N-terminal end of a hairpin module contribute to the structural stabilization of the protein (Hothorn et al. 2004a, b). In addition to these four conserved cysteine residues, the well conserved N-terminal end helical hairpin extension is crucial to the structural integrity of the protein. Moreover, the conserved C-terminal domain is thought to contribute to the interface stabilization of the protein (Hothorn et al. 2004a; Rausch and Greiner 2004). Amino acid motif PKF (proline, lysine and pheylalanine—orange highlight and boxed) in the α6 helix is highly conserved in all the sequences (Fig. 1). This motif is important in the complex formation between Arabidopsis cell wall/apoplastic invertase 1 (INV1) and a protein inhibitor (INH1) from tobacco (Hothorn et al. 2010) (Fig. 1). PKF in the cell wall/apoplastic invertase inhibitor has been shown to directly target the invertase active site (Lammens et al. 2008). In addition to PKF, the adjacent two amino acids, i.e. Gly (green shaded) and Ala (green shaded except in case of BrCIF/VIF, BnCIF/VIF, BoCIF/VIF, AtVIF1, GhCIF/VIF, TcCIF/VIF and BvCIF/VIF which is Gly instead) (Fig. 1) in α5 and α6 helices have also been shown to allow the insertion of consecutive loop residues into the INV1 active site cleft (Hothorn et al. 2010). Therefore, it can be hypothesized that the conservation of PKF in INH2 from potato might target the active site of the vacuolar invertase. However, the mechanism by which INH2 targets and regulates the vacuolar invertase activity is unknown. Therefore, it can be predicted that the non-synonymous SNPs of both the invertase inhibitors may have functional significance as the changes in amino acid sequence may alter the protein function and, as a consequence, the phenotype (Jehan and Lakhanpaul 2006). As INH1 and INH2 from potato shared high sequence identity at the amino acid level, both invertase inhibitors may participate in inhibiting invertase in potato tubers. However, studies on the interactions and inhibitory mechanisms between invertase and invertase inhibitor isoforms from potato are not reported.

Phylogenetic analysis and 3 D modelling of invertase and invertase inhibitory proteins

Invertase inhibitor protein sequences from planta (Table 2) were then subjected to phylogenetic analysis. The results presented on the phylogenetic tree are in agreement with other studies (Datir et al. 2012, 2019; Lin et al. 2013), in which two groups of invertase inhibitors corresponding to cell wall/apoplast and vacuole were evident (Fig. 2). The evolutionary conservation of this protein in potato is probably because the protein sequences of invertase inhibitors evolved slowly and originated from the common ancestor gene (Datir et al. 2012). Invertase inhibitor genes could have diverged as a gene duplication event because both INH1 and INH2 are located in tandem orientation on potato chromosome XII. INH2 is located 5.5 kb upstream with that of INH2 without any intervening genes (Brummell et al. 2011; The Potato Genome Sequencing Consortium 2011).

For 3D modelling, the protein sequences of cell wall/apoplastic invertase, vacuolar invertase INH1 and INH2 from potato (Table 1) were predicted (Phyre²—Tyagi et al. 2012) and the quality was further evaluated with PROCHECK (Laskowski 1993). The Ramachandran plots were generated for each structure (Supplementary Fig. S2). The results showed that the percentage of residues in favoured region and in allowed region along with Z score and QMEAN score (Table 4) indicated overall good, acceptable and reliable models (Benkert et al. 2009).

The cell wall/apoplastic invertase and vacuolar invertase sequences from potato (Table 1) were aligned for the comparison of conserved motifs (Fig. 3). A total five isoforms of cell wall/apoplastic invertase and one isoform of vacuolar invertase were considered in the present study. Plant invertase genes are known to contain six to eight exons (Alberto et al. 2004) with one conserved smallest functional exon (exon 2), which is only 9 bp long (Kim et al. 2000; Alberto et al. 2004). It encodes three amino acids (DPN), which are part of the highly conserved‚ β-fructosidase motif NDPN (Goetz and Roitsch 2000). This motif together with the well-conserved cysteine catalytic site (WECV/PDF) might have an important function in enzyme conformation or catalytic activity (Fotopoulos 2005). Cell wall/apoplastic invertase has a proline (PRO) residue in the catalytic site consisting of a cysteine residue, while the vacuolar invertase possesses a valine (VAL) residue instead (Goetz and Roitsch 2000). Cell wall/apoplastic invertase (Inv_ap-a and Inv_ap-b) and vacuolar invertase (Pain 1) isoforms of invertase are known to control CIS traits (tuber starch, tuber sugar and chip quality) in potato tubers (Menéndez et al. 2002; Li et al. 2008; Draffehn et al. 2010). Unlike cell wall/apoplastic invertase, only a single locus (chromosome III) has been reported for the vacuolar invertase gene in potato (Pain 1) (Menéndez et al. 2002; Draffehn et al. 2010). Potatoes possess five invertase genes, located on chromosomes III (Pain 1), IX (InvGE, InvGF and Inv_ap-b) and X (Inv_ap-a) (Chen et al. 2001; Draffehn et al. 2010). Inv_ap-a and Inv_ap-b are two cell wall / apoplastic bound isoforms of acid invertase and are located on chromosomes X and IX respectively (Chen et al. 2001). A putative intra-cellular soluble acid invertase is located on chromosome III (Pain 1). InvGE/GF are tandemly duplicated genes assigned to locus Inv_ap-b on chromosome IX (Menéndez et al. 2002), while InvCD141 and InvCD111 were assigned to Inv_ap-a locus on chromosome X (Li et al. 2008). A very high natural allelic variation using SNPs diversity noted for five invertase genes in potato was found to be associated with various tuber quality traits (Li et al. 2008). Despite these studies, the role of structural variants of invertase (cell wall/apoplastic and vacuolar), its inhibitors (INH1 and INH2) and their interactions in controlling the CIS traits in potato still remains to be elucidated.

In silico interactions between invertase and invertase inhibitors

To understand how different isoforms of invertases interact with invertase inhibitors, in silico interactions between different isoforms of invertases and invertase inhibitors were performed (Tables 5, 6; Fig. 4; Supplementary Fig. S4). In the complex structure, the interaction of LYS119 of INH1 with the acid–base catalyst GLU245/246 of cell wall invertase (Table 5) is in agreement with (Hothorn et al. 2010). In our studies the interaction of LYS119 of INH2 was found interacting with GLU398 of vacuolar invertase (Fig. 3; Table 6). Additionally, the adjacent conserved GLY116 (INH1) and GLY126 (INH2) (Fig. 4; Tables 5, 6) have been shown to allow the insertion of consecutive loop residues into INV1 active site cleft (Hothorn et al. 2010). ARG residue at the C terminus end is highly conserved in INH1 and INH2 (Fig. 4) (Hothorn et al. 2010). This residue was found to be interacting with ASP residue of invertase. However, in our studies, we noticed that ARG168 in INH1 is interacting with VAL281 (CWIInvGE), LEU277 (CWIInvGF), VAL279 (CWIpCD111) and VAL280 (CWIpCD141) while this ARG168 was contacting ASP334 of vacuolar invertase (Figs. 3, 4; Tables 5, 6). Surprisingly, we did not observe any interactions of PRO and PHE (Fig. 4, red boxed) with any residues of invertases. Other residues that are flanking the small insertion loop in INH1 (GLY116 and ALA121) and INH2 (GLY126 and ALA131) are highly conserved and favour a particularly sharp bent loop structure and allow the insertion of consecutive loop residues into the invertase active-site cleft (Fig. 4). Although no interactions were noted between conserved ALA (from INH1 and INH2) with any residue of invertase (cell wall/apoplastic invertase and vacuolar invertase isoforms), GLY residues established a contact with GLN and ARG residues of different cell wall/apoplastic invertase isoforms and ASP, GLN and ARG residues of vacuolar invertase (Tables 5, 6). Both GLN and ARG residues are highly conserved in both the invertases (cell wall/apoplastic invertase and vacuolar invertase) (Fig. 3).

The interactions between different isoforms of invertase and invertase inhibitors using modelling approaches suggested that, interestingly, INH1 displayed slightly more affinity with that of vacuolar invertase compared to INH2. On the other hand, INH2 showed a very strong interaction profile with the vacuolar invertase only (Tables 5, 6). Based on these observations, it can be concluded that both the inhibitors might be needed to inhibit the activity of invertase. The support for this hypothesis comes from the in vitro studies conducted by Sander et al. (1996), which demonstrated that the I_NT (apoplastic invertase inhibitor from tobacco) and I_LE (vacuolar invertase inhibitor from tomato) both inhibit cell wall/apoplastic invertase and vacuolar invertase. Furthermore, they found a much faster complex formation between tobacco INH1 and tobacco cell wall/apoplastic invertase than between tobacco INH1 and tomato vacuolar invertase. However, the mechanism of inactivation was clearly different for the cell wall/apoplastic invertase and the vacuolar invertase. Two invertase inhibitors (AtC/VIFs) from Arabidopsis thaliana revealed that the AtC/VIF1 showed specific inhibition of the vacuolar invertase activity, whereas the AtC/VIF2 inhibited both cell wall/apoplastic invertase and vacuolar invertase (Link et al. 2004). Interaction between the cell wall/apoplastic invertase and the INH1 was identified as potentially important in the CIS process (Baldwin et al. 2011). Furthermore, based on these interactions, they concluded that potato lines containing specific cell wall/apoplastic invertase and INH1 alleles were cold tolerant. Therefore, in silico predictions of interactions between invertase and invertase inhibitors can be utilized to identify the CIS-tolerant potato cultivars.

Previously, the allele diversity of both INH1 and INH2 genes along with their functional characterization was performed from the CIS resistant and susceptible potato cultivars (Brummell et al. 2011). The Potato Genome Sequencing Consortium (2011) data revealed that both INH1 and INH2 genes are tandemly duplicated on potato chromosome XII without any intervening genes. Moreover, it has been confirmed that the INH2 gene showed developmentally regulated alternative splicing. Furthermore, it has been shown that the splicing events resulted in INH2α transcript encoding the full-length protein and two other hybrid mRNAs such as INH2β*A and INH2β*B. These hybrids encoded deduced vacuolar invertase inhibitors with divergent C-termini. This was mainly due to mRNA splicing of an upstream region of INH2 to a downstream region of INH1 (Brummell et al. 2011). They further claimed that the increased amounts of invertase inhibitor in CIS resistant lines may contribute to the suppression of acid invertase activity and subsequently prevent the cleavage of sucrose. Increased RNA splicing activity was specifically observed in several potato lines resistant to CIS and therefore, these lines may generate a range of proteins with additional functional capacity to aid adaptability. These results revealed that both INH1 and INH2 proteins might be needed to control the activity of invertases in cold-stored potato tubers. Also, it can be speculated that the microheterogeneity, i.e. variation in the amino acid sequence of the protein observed in both of the inhibitors may target and control both invertases. Therefore, for the complete suppression of vacuolar invertase activity, both INH1 and INH2 might be required. However, both in vivo and in vitro studies are needed to affirm and support this hypothesis. The natural diversity of invertase and invertase inhibitors genes and the interactions of specific alleles of these genes during CIS traits might help in selection of superior genotypes in potato breeding programs (Li et al. 2008; Baldwin et al. 2011; Brummell et al. 2011). Studies on the interaction between these two proteins will help to manipulate the activity of invertase by selecting specific inhibitor alleles to improve the CIS performance of potato tubers using genetic engineering approaches.

Conclusions

Manipulation of invertases via their inhibitor proteins is necessary to overcome the problem of CIS in potato tubers. The complete inhibition of invertase is possible only when a sufficient amount of inhibitor is present in potato tubers under cold conditions. Our studies suggested that it is possible that both cell wall/apoplastic and vacuolar invertase inhibitors might be participating in the inhibition reaction with invertase. Therefore, understanding the structural properties of these two enzymes from different potato cultivars are crucial to get further insights into the functional aspects of invertase and their inhibitory proteins in the CIS process. This can be further explored by studying the combination of interactions between different allelic variations in invertase and invertase inhibitors from different CIS resistant and susceptible potato genotypes. The SNPs or the allelic variants of the cell wall/apoplastic invertase inhibitor and the vacuolar invertase inhibitor (Datir et al. 2012, 2019) might be responsible for the specific inhibitory reaction. It can be speculated that these variations might not bring any alterations in the structure and function, but may introduce alterations in specificity and inhibitory activity with invertase. In silico interactions might provide valuable information in selecting the proper combinations of invertase and invertase inhibitor from diverse potato genotypes. Then, the specific inhibitor alleles can be chosen from potato cultivar and used in the CIS potato breeding programs. Moreover, this may help researchers to design genetically engineered potatoes by selecting only specific alleles of the inhibitor for improved CIS performance.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Compliance with ethical standards

Conflict of interest

Authors declare that they have no conflict of interest.