Title: Independence and Interaction of Regions of the INNER NO OUTER Protein in Growth Control During Ovule Development

The outer integument of the Arabidopsis ovule develops asymmetrically with growth and cell division occurring primarily along the region of the ovule facing the base of the gynoecium (gynobasal). This process is altered in the mutants inner no outer ( ino ) and superman ( sup ), which lead to absent or symmetrical growth of the outer integument, respectively. INO encodes a member of the YABBY family of putative transcription factors and its expression is restricted to the gynobasal side of developing ovules via negative regulation by the transcription factor SUP. Other YABBY proteins (e. g. CRABS CLAW (CRC) and YABBY3 (YAB3)) can substitute for INO in promotion of integument growth, but do not respond to SUP regulation. In contrast, YABBY5 (YAB5) fails to promote integument growth. To separately investigate the growth-promotive effects of INO and its inhibition by SUP, domain swaps between INO and YAB3, YAB5, or CRC were assembled. The ability of chimeric YABBY proteins to respond to SUP restriction showed a quantitative response proportional to the amount of INO protein and was more dependent on C-terminal regions of INO. A different response was seen when examining growth promotion where the number and identity of regions of INO in chimeric YABBY proteins were not the primary influence on promotion of outer integument growth. Instead, promotion of growth required a coordination of features along the entire length of the INO protein, suggesting that intramolecular interactions between regions of INO may coordinately facilitate the intermolecular interactions necessary to promote formation of the outer integument.


INTRODUCTION
The YABBY family of genes participates in the specification of abaxial identity in plant lateral organs (Siegfried et al., 1999). INNER NO OUTER (INO) is an Arabidopsis YABBY family member that is specifically expressed in the ovule outer integument and is essential for ovule development (Baker et al., 1997;Villanueva et al., 1999). INO likely modulates the transcription of currently uncharacterized genes involved in the growth of the outer integument in Arabidopsis. While the molecular basis of INO function remains largely unknown, the ability of INO to interact with the putative transcription factor NOZZLE/SPOROCYTELESS (NZZ/SPL) (Sieber et al., 2004) further supports the hypothesis that INO modulates expression of genes necessary for outer integument development. Structural domains of YABBY members have been hypothesized based on sequence alignments that show similarities to known protein domain motifs (Siegfried et al., 1999;Villanueva et al., 1999;Bowman, 2000). A region toward the N-terminus is similar to Cys 2 -Cys 2 zinc finger domains and has been shown to interact with zinc ions in vitro for the YABBY protein FILAMENTOUS FLOWER (FIL) (Kanaya et al., 2001). A region (the "YABBY" domain) nearer to the C-terminus is predicted to contain two αhelical regions that are thought to participate in DNA binding, in large part due to sequence similarity with the DNA binding motif of the High Mobility Group (HMG) transcription factors (Sawa et al., 1999). An understanding of the functions of these regions and a more precise determination of the boundaries of each putative domain would facilitate our understanding of how INO mediates outer integument growth in Arabidopsis.
During Arabidopsis ovule development, outer integument growth is restricted primarily to the side of the ovule that faces the basal region of the gynoecium (gynobasal) (Fig. 1). This asymmetry is mediated via SUPERMAN (SUP), a transcription factor that has been shown to restrict INO expression to the gynobasal side of the developing ovule (Meister et al., 2002).
While ino mutants failed to initiate or support outer integument growth, it was found that sup  (Meister et al., 2002). To further test the nature of INO and SUP antagonism, Meister et al. (2002) (Meister et al., 2005). While all CRC/INO domain swap lines were able to support outer integument growth, only a subset responded to SUP repression of growth on the gynoapical side of ovules. This suggested that all three regions of INO participate in both the promotion of growth and response to SUP inhibition, although the combination of regions 3 and 4 appeared to play a more significant role (Meister et al., 2005).
These results, however, may have been specific to domain swaps between INO and CRC due to the unique role of CRC in gynoecium and nectary specification, in addition to its role in abaxial identity that is common among all YABBY proteins (Alvarez and Smyth, 1999). This is particularly of concern due to the possibility that INO and CRC regions may share some redundancy in function that is not present in other YABBY members as INO and CRC are the only members with specific reproductive roles.

Nomenclature
In the following sections, representation of the source of each of the exchanged regions in the chimeric cDNAs is indicated by the following: "I" indicates INO sequence; "C" indicates CRC sequence; "3" indicates YAB3 sequence; and "5" indicates YAB5 sequence. These designations are listed from left to right starting from the amino terminus and ending with the carboxy terminus. The majority of exchanges focused on three major regions of INO: the Nterminal and Zn-finger regions (1) comprised the first region, the central variable region (2) comprised the second region, and the YABBY (3) and C-terminal (4) regions together comprised the third region ( Fig. 2A). However, since the third region possessed greater INO-specific function than the first and second regions in a prior study (Meister et al., 2005), we separately exchanged the YABBY (3) and C-terminal (4) regions in a subset of our experiments. Separate exchange of the C-terminal region 4 is indicated by the following: "i" indicates INO sequence; "3" indicates YAB3 sequence; and "c" indicates CRC sequence. Boundaries between regions were selected based on sequence conservation between YABBY members as well as results indicating the importance of seven conserved residues at the carboxyl end of the Zn-finger region ( Fig. 2B) (Siegfried et al., 1999;Villanueva et al., 1999;Bowman, 2000;Meister et al., 2005).

INO/YAB3 Chimera Expression Using P-INO
Chimeric cDNAs composed of INO and YAB3 sequences were expressed using PRO INO and phenotypic effects on outer integument development were scored in an ino-1 mutant background (Table I). Five phenotypic classes were observed among transformed lines containing chimeric transgenes: sup-like, weak-sup, wild type, weak-ino, and ino-like (Fig. 3). In order to determine the significance of differences observed in ovule development for chimeric lines as compared to the control lines, the data were subjected to a Fisher's Exact test.
A Bonferroni adjustment was used to adjust the alpha value to account for the increased probability of error when numerous pairwise comparisons are made within a single set of data (Sonnenberg, 1985). Statistical analyses of all domain swap lines in comparison to control lines are reported in the supplementary material (Tables S1-S6).  (Tables I, II). Both I33 and I3I chimeras were significantly different from INO and 3II, but were not significantly different from II3. This indicates that adjacent regions 2, 3, and 4 are better able to support growth like that of INO than adjacent regions 1, 2, and 3 ( Fig. 2A able to support growth of the outer integument, the extent of growth varied among lines (Table   I). Based on our results, the central variable region of YAB3 contains sufficient ability to promote some growth of the outer integument on the gynoapical side of ovules, as evidenced by PRO INO :I33, I3I, and 33I transgenic lines. The presence of adjacent regions from YAB3 did not enhance the ability of chimeric proteins to stimulate outer integument growth, since PRO INO :I33 and 33I were not significantly different from PRO INO :I3I (Table II). The results for PRO INO :333i indicate that the C-terminus of YAB3 is not necessary for the promotion of growth nor escape from SUP repression when expressed in ovules, as these transgenics were not significantly different from PRO INO :333 plants. This also shows that all three primary regions of INO contribute to its ability to respond to SUP, with the central variable region providing the smallest increment of this activity.

INO/YAB3 Chimera Expression Using PRO 35SCaMV
In parallel with phenotypic analysis of chimeric YABBY expression using PRO INO , ectopic expression using the generally active Cauliflower Mosaic Virus 35S promoter (PRO 35SCaMV ) was pursued in order to address the functionality of the chimeric cDNAs. When expressed ectopically, YABBY family members, including both YAB3 and INO, have been shown to alter leaf morphology leading to narrowed and curled leaves likely due to the abaxialization of adaxial tissue types (Eshed et al., 1999;Siegfried et al., 1999;Meister et al., 2005).

INO/YAB5 Chimera Expression Using PRO INO
The majority of ino-1 mutant plants harboring the PRO INO :555 (YAB5) construct exhibited no growth of the outer integument, and when any growth was observed it was very limited (Table I) Table I). The data were analyzed as for the YAB3/INO domain swaps.

PRO INO :III Versus PRO INO :INO/YAB5 Swaps
Although the ability to support outer integument growth was markedly less than that seen for INO/YAB3 domain swaps, all INO/YAB5 chimeras were able to support some degree of growth (Table I) (Table I). This is largely because a YAB5/INO chimeric gene containing any one region of INO was sufficient to support some growth of the outer integument for all chimeric lines observed. As expected, ino-1 plants were not significantly different from PRO INO :555 plants, since YAB5 was rarely able to support any growth of the outer integument (Table I).

INO-YAB5 Chimera Expression Using PRO 35SCaMV
As with ectopic expression analysis of YAB3 and INO domain swaps, all chimeric lines of domain swaps between YAB5 and INO resulted in the characteristic YABBY overexpression phenotypes observed previously (Fig. 4) (Eshed et al., 1999;Siegfried et al., 1999;Meister et al., 2005). Thus the weak or absent ability of the chimeric proteins to support integument growth results from specific properties of the proteins, rather than being an indication of a generally inactive protein.

C-terminal Domain Swaps Using PRO INO
As mentioned previously, domain swap experiments that exchanged regions 1, 2 or 3 and 4 from CRC and INO suggested that regions 3 and 4 provided more functional information than regions 1 and 2 ( Fig. 2A) (Meister et al., 2005). In order to address whether this specific information was contained within region 3 or 4, these regions were swapped to individually evaluate their roles. When the C-terminal domain swaps were expressed using PRO INO , the majority of lines showed an ability to support outer integument growth (Fig. 3). The strategy for classification described above was used to evaluate phenotypes and statistical significance.

C-terminal Domain Swaps Using PRO 35SCaMV
As for the other tested constructs, the C-terminal domain swaps were tested for YABBY function through expression from PRO 35SCaMV . The resulting chimeric lines exhibited leaf abnormalities consistent with ectopic YABBY effects. Thus, the limited function of CCCi in supporting integument growth does not appear to result from a general loss of all protein activity. DNA targets, but fails to drive proper expression of these genes. This is supported by our observation that 5I5 chimeras, which contain DNA-binding domains exclusively of YAB5 origin, supported outer integument growth to a limited extent, although we cannot rule out the potential role of the central variable region in this interaction.

DISCUSSION
It is likely that the binding of YAB5 to INO DNA targets, or its interaction with transfactors needed for expression of these targets, is inefficient due to the absence or misorientation of critical residues that facilitate these interactions. Reduction in these interactions could be

C-terminal Domain Swaps
YAB3/INO C-terminal domain swaps had little effect on the ability of the chimeric proteins to influence ovule development (Table I). These results suggest that the C-terminal portion of INO does not include information that can overcome the sup-like response observed when the YAB3 protein is expressed in ovules.
In contrast with the C-terminal domain swaps between YAB3 and INO, the swaps between CRC and INO show that the C-terminal region of CRC does possess unique information ( Table I). The distribution of phenotypes among transgenics with CCCi spans the entire range of phenotypic classes; however, the majority of plants possessed weak-ino ovules. This suggests that the ability of the CRC protein to both support outer integument growth and overcome SUP repression is at least partially dependent on certain residues contained within the C-terminal region. This could reflect the need for specific intermolecular interactions between the Cterminal region and the rest of the CRC protein for formation of an active structure. In further support of the importance of the C-terminal region, while most plants containing PRO INO :IIIc exhibited a wild-type ovule phenotype, several transgenic plants possessed weak-sup ovules and sup-like ovules (Table I). Thus, the presence of the C-terminal portion of CRC was sufficient for the IIIc protein to overcome SUP repression in some plants. This ability, however, is not exclusively found within the C-terminal region as PRO INO

CONCLUSIONS
We have found that no particular region of INO contained the specific information responsible for either the differential promotion of outer integument growth or differential response to SUP. This parallels results described for INO/CRC domain swaps in Meister et al.

Construct Asssembly
Chimeric coding sequences: Chimeric coding sequences were created using overlap extension PCR (Horton et al., 1990)  Assembly of these constructs was performed using the restriction sites BamHI/XbaI or BglII/XbaI according to the particular chimeric cDNA used (data not shown). This cloning strategy replaced the CRC coding sequence originally incorporated in pRJM42.

Plant Growth and Transformation
Fragments including the chimeric cDNAs in the appropriate expression cassettes listed above were excised with NotI, the chimeric genes were inserted into pMLBART using the same site, and the resulting plasmids were transferred into the Agrobacterium strain ASE via triparental matings (Figurski and Helinski, 1979;Fraley et al., 1985;Gleave, 1992 1987;Clough and Bent, 1998). T1 lines were selected following germination on soil by treatment with the herbicide glufosinate10-ammonium (Finale™, Farnham Companies, Phoenix, AZ) which selects for the phosphoinothricine acetyl transferase gene. Homozygous ino-1 plants were identified using a detectable polymorphism between Col and Ler ecotypes at the INO locus as described previously (Meister et al., 2002).

Microscopy
Determination of phenotype class for transformants containing chimeric cDNAs expressed using PRO INO was initially performed using dark-field microscopy. Representative lines of each phenotypic class were then fixed and prepared for analysis using Scanning Electron Microscopy following methods described previously (Broadhvest et al., 2000).

Statistical Analysis
Statistical analysis of significant differences between data sets of each chimera tested and against control lines was performed using Fisher's Exact Test at the UC Davis Statistical Laboratory. Significant differences between phenotypic effects of the transgenes were calculated with the modified Bonferroni adjustment (alpha-value/number of pairwise comparisons) in order to reduce the rate of false-positives while conducting simultaneous pairwise comparisons. This method has been shown to be advantageous when multiple statistical analyses are conducted simultaneously (Wright, 1992;Morikawa et al., 1996).

Figure 1.
Scanning electron micrographs of wild-type and mutant ovules. During wild-type ovule development, primordia emergence is followed by the appearance of distinct structures including the funiculus, outer and inner integuments, and the nucellus (A). Asymmetric growth is observed as the ovule develops (B) until the micropyle lies adjacent to the funiculus in mature ovules (C, D). In ino-1 ovules (E), the outer integument fails to initiate and as a result, the symmetrical growth of the inner integument and nucellus are apparent. In sup-5 ovules (F), the outer integument exhibits growth on both sides, leading towards a concentric ring of symmetrical growth. c, chalaza; i, inner integument; m, micropyle; n, nucellus; o, outer integument. For all panels, the gynobasal side is at left. Scale bar = 25µm. The Zn-finger and YABBY regions, the borders of which were determined based on previous sequence alignments (Siegfried et al., 1999), are indicated using single solid and broken underlines, respectively. The seven conserved amino acid residues adjacent to the previously defined zinc finger motif are indicated with a double underline. Residues on either side of protein region boundaries that were used in assembly of chimeric proteins are highlighted in black.