Hiding in plain sight: the F segment and other conserved features of seed plant SKn dehydrins

Main conclusion An 11-residue amino acid sequence, DRGLFDFLGKK, is highly conserved in a subset of dehydrins found across the full spectrum of seed plants and here given the name F-segment. An 11-residue amino acid sequence, DRGLFDFLGKK, is highly conserved in identity and polarity in 130 non-redundant dehydrin sequences representing conifers and all major angiosperm groups. This newly described motif is here given the name F segment based on the pair of hydrophobic F residues at the core of the sequence. The majority of dehydrins previously classified as SKn dehydrins contain one F segment N terminal to the S and K segments and can accordingly be reclassified as FSKn dehydrins. A cysteine-containing variant, GCGMFDFLKK, occurs in a few rosid and asterid taxa. The S segment in this and other dehydrin types also includes previously overlooked conserved features, including a KLHR prefix and charged or G residues within and following the characteristic string of S residues. Secondary structure prediction models indicate that the F segment and S segment prefix may form amphipathic helices that could be involved in membrane or protein binding.


Introduction
Dehydrins are a family of land plant proteins that may be expressed constitutively at low levels but are often produced de novo or at higher levels in response to drought, low temperature, or other stresses. They have been detected by western blotting or nucleic acid sequencing in all types of land plants, including bryophytes, pterophytes, gymnosperms, and all major groups of angiosperms. The main structural features of dehydrins were first described over 20 years ago (Close 1996), and include the number and modular arrangement of three types of short, distinctive segments in the protein, designated K, Y, and S segments. Using the YSK shorthand nomenclature proposed by Close, most known dehydrins fall into five types: K n , SK n , K n S, Y n K n , and Y n SK n ,.
With a very few exceptions (noted below) the common feature of all dehydrins is one or more repeats of the K segment, a lysine-rich, 15 amino acid sequence with a highly conserved pattern of charged and nonpolar residues. The consensus K segment sequence in angiosperms is EKK-DIMGKIKEKLPG, with a generally similar pattern conserved in conifer variants (Jarvis et al. 1996;Perdiguero et al. 2014). The identity or polarity of all K segment residues is highly conserved. The K segment forms an amphipathic ahelix in nonpolar environments and binds membranes (Koag et al. 2003(Koag et al. , 2009, suggesting that a primary function of dehydrins is to protect membranes and perhaps proteins against dehydration stress. The Y segment comprises seven residues with the consensus sequence (V/T)DEYGNP. When present, usually one to three closely spaced copies of the Y segment are located towards the N terminal end of the protein relative to the S or K segments. A group of four Cornus sericea dehydrins contain from 14 to 35 recognizable Y segments & G. Richard Strimbeck richard.strimbeck@ntnu.no (Sarnighausen et al. 2004), and there are a few dehydrins with four. Unlike the K and Y segments, as a rule the S segment occurs only once in any one dehydrin. It is recognized as a series of from three to nine consecutive serine residues, often interrupted by a single, usually polar residue, often D or E, after the first S. It is located N-terminal to the first K segment or at the C-terminus. Dehydrins with S segments have been found in all major seed plant taxa and the moss Physcomitrella patens. These segment types are bracketed and separated by unconserved regions, sometimes called phi segments. These can vary in length from a few to over 100 residues, and are rich in glycine and polar residues, so that dehydrins as a group are highly hydrophilic.
These basic features are found in the several hundred complete dehydrin sequences in online databases, and have been discussed and reaffirmed in numerous review articles on dehydrins (Close 1996;Allagulova et al. 2003;Rorat 2006;. There are only a few known exceptions to the five dehydrin groups described by Close (1996), for example a K 4 SK 2 dehydrin in Picea abies (GenBank accession ABS58630.1), an SK 3 S variant in Stellaria longipes (CAA79709.1), and a Y 2 KY 2 KY dehyrdrin in Juglans regia (AGJ94410.1). Also only few proteins with recognizable Y or S segments and sequence similarity to other dehydrins but with truncated or completely lacking K segments are described. These include a group of conifer proteins (Perdiguero et al. 2014), a Y 14 S protein in C. sericea (Sarnighausen et al. 2004), and complete ORF sequences with terminal S segments but no clearly identifiable K segments in the Prunus persica (XP 007202808) and Solanum lycopersicum (XP 010317810.1) genomes. Perdiguero et al. (2012) recently reported on two additional conserved segments in Pinaceae SK n dehydrins.
While compiling a library of dehydrin amino acid sequences, I noticed a short, palindromic sequence in many SK n -type dehydrins: GLFDFLG, located near the N-terminal end of the protein. Further inspection suggested a longer conserved sequence with the preliminary 11 residue consensus DRGLFDFLGKK. Here I report that the majority of SK n dehydrins contain a single copy of this segment, which I have named the F segment, and can thus be designated FSK n dehydrins. I decribe its taxonomic breadth and sequence conservation and variation. I also note additional sequence conservation around the S segment in both FSK n and Y n SK n dehydrins.

Materials and methods
To compile a library of dehydrin sequences, I searched the NCBI data for sequences identified as dehydrins and conducted BLAST searches on the angiosperm and conifer consensus K segment sequences. The results were screened to identify and remove sequences lacking core dehydrin characteristics, allelic variants, and highly similar homologs within taxa. Genera such as Pinus, Vitis, and Solanum are overrepresented in NCBI by registration of dozens of allelic variants and nearly identical homologous sequences in closely related taxa. Based on the number and occurrence of K, Y, and S segments, the resulting 395 sequences were classified into the five types described by Close, with the few exceptions noted. It was in this process that I first noticed the F segment.
After a BLAST search in the NCBI database for proteins containing variations of the 11-residue sequence DRGLFDFLGKK, I screened the results as above to remove non-dehydrins and redundant variants within taxa. Sequences were aligned using Clustal Omega and the alignments checked and adjusted manually. To explore sequence variation in and around the conserved segment, I isolated and aligned a 32-residue sequence centered on it and tallied the amino acids of each type found at each position.
To explore variation in and around the S segment, all S-segment containing dehydrins from the library were extracted. I excerpted a 35-residue segment centered on the consecutive S residues. These excerpts were manually aligned using the first S residue, recognizable variants of a commonly occurring HR prefix, and the first occurrence of a charged residue, usually D or E, following the S residues. Alignment of the latter was forced by inserting blanks in sequences with fewer than the maximum number of nine S residues. Following alignment, I tallied the amino acids at each position for each of the four dehydrin types (including FSK n ) that incorporate S segments.
The eight-structure class protein secondary structure prediction model SSpro8 (Magnan and Baldi 2014) was run on a selection of full-length FSK n dehydrin sequences to assess the potential for helical or sheet structures in and around the F and S segments.

Frequency of the F-segment in dehydrins
In the initial screening and classification of 395 minimally redundant dehydrin sequences recovered from NCBI, 93 were classified as SK n type dehydrins based on the presence of S and K but no Y segments (Table 1). I found F segments in 82 of these, indicating that most known SK n type dehydrins can be reclassified as the FSK n type. There were only nine FK n type dehydrins in the sample, and only one with more than one copy of the F segment, an F 3 SK 2 dehydrin in Rhododendron (AGI36547.1). There were no sequences with both Y and F segments. An additional 75 sequences that were incomplete at the N terminal end could potentially be any of a number of types including FK n or FSK n type dehydrins. About 25% of the 72 complete gymnosperm dehydrin sequences in the library contained F segments, but none contained Y segments (Table 1). The absence of Y segments in gymnosperms was noted by Perdiguero et al. (2012Perdiguero et al. ( , 2014, but not in earlier reviews of dehydrin occurrence, structure, and function (e.g. Close 1996; Allagulova et al. 2003;Rorat 2006;. There were also no complete gymnosperm SK n or K n S dehydrins in the curated library, so that the available data indicates that in gymnosperms the S segment is always preceded by an F segment.

Conservation variation in the F segment
The BLAST search of the NCBI database for proteins containing variations of the 11-residue sequence DRGLFDFLGKK yielded an initial library of about 575 complete and partial sequences, about 300 of which were readily recognized as dehydrins based on the presence of one or more recognizable variants of the K segment. After screening out redundant sequences, I arrived at a list of 130 minimally redundant proteins representing a broad spectrum of seed plant groups, including conifers as representative of gymnosperms and a variety of groups within the monocot, caryophyllid, rosid, and asterid clades of angiosperms. 123 of these also contained an S segment.
The 11-residue DRGLFDFLGKK sequence specified in the BLAST search is well-conserved across all 130 sequences, with the core 6-residue sequence GLFDFL nearly invariant in identity or polarity in all variants across the taxonomic spectrum (Fig. 1). The G following this core sequence is deleted in about 40% of the sequences, and there is an apparent insertion of a second hydrophobic residue after the conserved F/L in position 8 in Poaceae. In a subset of 12 rosid sequences and two asterid sequences, the initial DR is replaced by GC and the second G is omitted, giving the consensus GCGMFDFLKK in this variant. The conservation of a rare and metabolically expensive cysteine residue suggests some additional functionality associated with this couplet. The terminal K residues, typically followed by three to five more charged residues, can be interpreted as general features of dehydrins but may also contribute to the functionality of the segment. Similarly, a well-conserved E residue at position -3 may contribute some functionality.
The F segment seems to have gone unnoticed over the ca. 25 years of investigation into dehydrin structure and function. To give credit where it is due, Perdiguero et al. (2012) noted a 23-residue N-terminal sequence that includes the F segment and appeared to be conserved in conifers and, with considerable indel variation, in a few angiosperms. My analysis indicates that the core F segment is highly conserved in seed plants, and I propose that it should be recognized as a fourth functional segment type in the dehydrin family.

Conserved features of the S segment
The S segment has been previously characterized as a string of three to as many as 13 serine residues, and typically followed by a string of charged K, D, or E residues . In K n S dehydrins it is located at the C-terminus of the protein. In other types it is N-terminal to the first K segment and in between it and F or Y segments where these are present, as indicated by nomenclature.
Additional conserved features around the S segment emerge on closer examination. In dehydrins with S segments on the N terminal side, the first S in the sequence is typically followed by a D, E, or G residue, which is then followed by two to as many as 12 additional serine residues. In K n S dehydrins, this is mirrored by a highly conserved terminal DSD motif, which, in the 25 K n S sequences in my curated sample, always forms the C-terminus of the protein. In FSK n dehydrins, there is a well-conserved KLHR prefix (Fig. 2), and the C terminal S is followed by 2-7 acidic residues, giving the generalized sequence KLHRS(D/E)S 2-12 (D/E) 2-7 , followed by a string of predominantly G and charged residues. The prefix is somewhat less conserved in Y n SK n and SK n dehydrins, where the initial K is not conserved and a polar residue may replace H. K n S dehydrin S segments have a more variable mix of charged, H, and G residues as a prefix.

Structural modeling
Predicted structures from SSpro8 (Magnan and Baldi 2014) for a selection of FSK n dehydrins consistently suggest a short helical region centered on the five residue core sequence LFDFL in the consensus F segment and the GCG variant, including those where the D is replaced by a G residue. In a helical wheel projection of this short sequence, the four hydrophobic residues are arrayed on one side of the helix, with the charged D or somewhat hydrophilic G residue opposing them. This observation suggests that the F segment could have amphiphilic membrane or protein binding properties similar to those of the K segment. The same model also consistently predicts helical regions N terminal to the S segment. The conserved pattern of hydrophobic residues in this region (Fig. 2) could result in an amphiphilic structure here as well. All K segments are also consistently helical in the SSpro8 predictions.

Conservation implies function
The high degree of conservation in the K, Y, F, and S segments in these otherwise highly variable and unstructured proteins implies that all four segment types play important roles in the overall function of the protein.
Functional studies show that that the K segment binds membranes and therefore suggest that the main function of dehydrins is membrane protection (Koag et al. 2003(Koag et al. , 2009). Membrane binding may be modulated by phosphorylation of S and pH-dependent dissociation of H residues , suggesting a function for the conserved H and S residues in the S segment. The functions of the Y segment remain a mystery. Close et al. (1993) found some similarity between the Y segment and a nucleotide binding site in bacterial chaperones, an observation that has been echoed in some reviews (e.g. Allagulova et al. 2003;Rorat 2006). That similarity is not at all obvious on inspection of the sequences in the original report (Martin et al. 1993), and in any case there seems to have been no follow-up on this assertion. Whatever its function, gymnosperms are able to survive in a wide range of environments, including extremely dry or cold conditions, without any apparent need for the Y segment.
As noted above, the F segment may form a short, amphipathic helix with membrane or protein binding properties, and the S segment prefix may also have amphipathic properties. Sucrose, raffinose, and various compatible solutes have been hypothesized to stabilize (Carpenter and Crowe 1988) or replace (Clegg 1985) hydration shells around proteins or membranes; perhaps dehydrins play a similar role. Alternatively, binding by K segments or other amphipathic regions could anchor dehydrins to membranes or proteins so that they could act as ''molecular spacers'' (Strimbeck et al. 2015), preventing close approach and conformational or phase changes associated with repulsive forces under dehydration (Wolfe and Bryant 1999). Protein modification and binding studies (e.g. Koag et al. 2003Koag et al. , 2009 or other protein modification experiments may help clarify the binding and related protective properties of the F segment and other conserved regions of the different dehydrin types.
Author contribution statement GRS conducted all of the original analyses reported in this article.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://crea tivecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Fig. 2 Sequence conservation and variation in S segments from 82 FSK n dehydrins. p polar, h hydrophobic, plus or minus sign charged, minus sign acidic, plus sign basic, asterisk no consensus