Identification by Chimera Formation and Site-selected Mutagenesis of a Key Amino Acid Residue Involved in Determining Stereospecificity of Guinea Pig 3-Hydroxysteroid Sulfotransferase Isoforms*

The 3-hydroxysteroid sulfotransferases that have been isolated and cloned from humans and rodents appear to have broad substrate specificities. In the guinea pig, however, two 3-hydroxysteroid sulfotransferases have been isolated that function according to an innate stereospecificity: the α-isoform acts on steroids with a 3-hydroxyl group oriented in the α position, whereas the β-isoform acts on steroids where the 3-hydroxyl group is in a β orientation. To examine the structural basis for this remarkable stereoselectivity, chimeras of the two enzymes, which are 87% identical, were constructed. A chimera consisting of the NH2-terminal 91 amino acids of the α-isoform and the COOH-terminal 196 amino acids of the β-isoform displayed activity similar to that of the α-isoform. Site-selected mutagenesis of this 3α/β-hydroxysteroid sulfotransferase chimera involving the 12 amino acid differences that exist between the two isoforms within the 91 amino acid NH2-terminal region revealed that the amino acid residue at position 51 plays a fundamental role in determining the stereospecificity exhibited by the α- and β-isoforms,i.e. if residue 51 is an asparagine, α activity predominates, whereas if an isoleucine is in that position, β activity prevails.

The 3-hydroxysteroid sulfotransferases that have been isolated and cloned from humans and rodents appear to have broad substrate specificities. In the guinea pig, however, two 3-hydroxysteroid sulfotransferases have been isolated that function according to an innate stereospecificity: the ␣-isoform acts on steroids with a 3-hydroxyl group oriented in the ␣ position, whereas the ␤-isoform acts on steroids where the 3-hydroxyl group is in a ␤ orientation. To examine the structural basis for this remarkable stereoselectivity, chimeras of the two enzymes, which are 87% identical, were constructed. A chimera consisting of the NH 2 -terminal 91 amino acids of the ␣-isoform and the COOH-terminal 196 amino acids of the ␤-isoform displayed activity similar to that of the ␣-isoform. Site-selected mutagenesis of this 3␣/␤-hydroxysteroid sulfotransferase chimera involving the 12 amino acid differences that exist between the two isoforms within the 91 amino acid NH 2 -terminal region revealed that the amino acid residue at position 51 plays a fundamental role in determining the stereospecificity exhibited by the ␣and ␤-isoforms, i.e. if residue 51 is an asparagine, ␣ activity predominates, whereas if an isoleucine is in that position, ␤ activity prevails.
A major metabolic route in the biotransformation of steroidal compounds is via sulfoconjugation or sulfonation (1). Steroid sulfonation, which occurs widely across species (2), is catalyzed by a family of enzymes termed sulfotransferases, enzymes typically isolated from the cytosolic fraction of tissue preparations (3). Sulfotransferase enzymes are of immense biological significance in that they are involved in both genomic and nongenomic steroid action (4).
Steroid sulfotransferases acting on neutral steroids have been isolated and cloned from several species (4); these enzymes, although not always carefully examined, appear for the most part to have broad substrate specificities. An apparent exception to this rule was thought to be the guinea pig; in this species, two physically distinct forms of 3-hydroxysteroid sulfotransferase (HST), 1 specifically a 3␣-HST and a 3␤-HST, were isolated from the adrenal gland (5). The guinea pig 3␣-HST and 3␤-HST enzymes demonstrated remarkable substrate selectivity with respect to the spatial orientation of the ring A 3-hydroxyl group (5). Furthermore, these enzymes were considered to be not only stereoselective but also site-specific in that they acted only on a 3-hydroxyl group. It was recently discovered, however, that the 3␣-HST isoform, in contrast to the 3␤-HST isoform, would also sulfonate the 17␤-hydroxyl group of testosterone 2 and estradiol (6). Nevertheless, with respect to the 3-hydroxyl group, the 3␣-HST enzyme acts only on 3␣-hydroxylated steroids such as androsterone and allopregnanolone, whereas the 3␤-HST enzyme acts on 3␤-hydroxylated steroids such as pregnenolone and dehydroepiandrosterone. Thus, regarding the 3-carbon position of neutral steroids, the guinea pig HST isoforms appear to behave quite differently from their rodent and human counterparts by exhibiting a more precise substrate stereospecificity. The 3␣-HST isoform with a calculated molecular mass of 33,637 daltons and pI of 6.3 (7) and the 3␤-HST isoform with a calculated molecular mass of 34,033 daltons and pI of 7.3 (8) have been cloned and kinetically analyzed.
The structural basis for the stereoselectivity of the 3␣-HST and 3␤-HST isoforms, which are 87% identical, is not known. Therefore, we set out to examine this question by creating chimeras of the two 3-HST isoforms in combination with siteselected mutagenesis. The results of these studies, which form the basis for this report, suggest that the 3-hydroxy stereospecificity shown by guinea pig 3␣-HST and 3␤-HST largely resides within the NH 2 -terminal 91 amino acids of the proteins. In particular, the amino acid residues at position number 51 plays a crucial role.

Construction of Wild Type and Chimeric cDNAs
Full-length cDNAs encoding the 3␣-HST and 3␤-HST isoforms were subcloned into the bacterial expression vector pProEX HTc (Life Technologies, Inc.) at SalI and NotI sites. Briefly, the HST cDNAs (nucleotides Ϫ7 to ϩ896) were amplified by PCR using primers designed to contain SalI and NotI sense primer (5Ј-ACGACGTCGACCACCACCAT-GTCAGATAACACTC-3Ј) and antisense primer (5Ј-ATAACAATTGCG-GCCGCCACAGCAGTATCCAAATAAC-3Ј) restriction enzyme sites. Amplified DNAs were purified by Wizard PCR Preps purification system (Promega, Madison, WI), digested with SalI and NotI, and purified by 1% agarose gel electrophoresis followed by Gene Clean II (Bio 101 Inc., Vista, CA). The restricted DNAs were ligated into the expression vector that was similarly digested and purified. The resulting plasmids (HST-pProEX HTc) were amplified using Max Efficiency DH5␣ compe-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Chimeras were constructed utilizing a unique EcoO109I restriction enzyme site present at nucleotide 273 in both the 3␣-HST and 3␤-HST cDNAs. The HST-pProEX HTc plasmids were digested with SalI and EcoO109I, and the restriction fragments were separated by 1.2% agarose gel electrophoresis and purified by Gene Clean II. To construct the 3␣/␤-HST chimera, the small SalI/EcoO109I restriction fragment of the 3␣-HST upstream cDNA was ligated to the large SalI/EcoO109I restriction fragment corresponding to the 3Ј portion of 3␤-HST as well as the vector; the 3␤/␣-HST chimera was similarly constructed.

Site-selected Mutagenesis
Site-selected mutagenesis was performed by PCR using Pfu DNA polymerase according to the method of Quick Change site-directed mutagenesis (Stratagene, La Jolla, CA). Based on the amino acid sequence alignment of 3␣-HST and 3␤-HST, primers were designed to make amino acid substitutions at residue sites where a difference existed between the two isomers. In all, eight constructs were prepared as outlined in Table I; primers used to produce the eight constructs are depicted in Table II.
Construct Numbers 1-6 Involved the 3␣/␤-HST Chimera (Table I)-Each construct was produced using the appropriate primers (as denoted in Table II) and template (as indicated below). The aim here was to substitute ␤-HST-specific amino acid residues for ␣-HST-specific amino acid residues in the ␣-HST portion of the 3␣/␤-HST chimera. For construct number 1, the codons for amino acids Cys 17 and Ile-22 in the chimera were mutated to yield Gly and Leu residues using the 3␣/␤-HST chimera as template. For construct number 2, in addition to the mutations carried out for construct number 1, the codons for amino acids Phe-80, Thr-81, Gly-82, and Tyr-84 were altered to generate Thr, Gln, His, and Asn residues at those respective sites using construct number 1 as template. For construct numbers 3-6, construct number 2 was used as a template; in construct number 3, the codon for Pro-6 was changed to produce a Leu; in construct number 4, codons for Leu-65 and Asn-71 were mutated to yield Trp and Ile, respectively, at those sites; in construct number 5, codons for Val-87 and Gly-89 were altered to create Met and Ser, respectively, at those sites; in construct number 6, the codon for Asn-51 was changed to produce an Ile at that site. (Table I)-Each of these constructs was produced using the appropriate primers as detailed in Table II and 3␤-HST wild type as template. For construct number 7, the codon for Ile-51 was altered to create an Asn at that position; for construct number 8, the codon for Ile-71 was mutated to produce an Asn at that site.

3-Hydroxysteroid Sulfotransferase Expression and Purification
The HST chimeras and mutant constructs were used to transform Max Efficiency DH5␣ competent bacteria (Life Technologies, Inc.). Positive colonies were selected by PCR using HST-specific primers and bacteria in the PCR reaction mixture. Plasmids were purified from positive colonies and sequenced using Thermo Sequenase Radiolabeled Terminator Cycle Sequencing kit (Amersham Pharmacia Biotech). The selected bacteria were grown at 32°C to an A 600 of 0.5 at which time 1 mM isopropyl-1-thio-␤-D-galactopyranoside was added to induce fusion protein synthesis, and the incubations were continued for an additional 4 h. Bacterial cells were harvested by centrifugation and lysed in 1:1 (w/v) of column buffer (50 mM soduim phosphate, 300 mM NaCl, pH 8.0) containing 10 mM imidazole and 1 mg/ml lysozyme followed by brief sonication. Bacterial lysates were centrifuged at 100,000 ϫ g for 1 h at 4°C, and the supernatants were applied to nickel affinity columns (Qiagene, Inc., Valencia, CA). Columns were washed with 20 mM imidazole in column buffer, and the bound fusion proteins were eluted with 250 mM imidazole in column buffer. Eluted proteins were concentrated and equilibrated with 20 mM Tris-HCl, pH 7.5, containing 15% glycerol using a Centriprep 30 (Amicon, Inc., Beverly, MA). To remove the fused protein from HST, the concentrated fusion protein preparation (0.1 mg) was incubated with 500 units of recombinant TEV protease (Life Technologies, Inc.) for 16 h at 4°C and resubjected to nickel affinity column chromatography. Unbound protein in the flow-through was collected,   Fig. 2B). Location of amino acid substitutions are indicated for each numbered construct of the 3␣/␤-HST chimera (A) and 3␤-hydroxysteroid sulfotransferase (3␤-HST) wild type (B). For substitutions involving the 3␣/␤-HST chimera, ␣-HST amino acid residue substitutions precede the residue numbers, whereas the ␤-HST amino acid substitutions follow the residue numbers. Conversely, for the 3␤-HST wild type enzyme, ␤-HST amino acid residue substitutions precede the residue numbers, whereas the ␣-HST amino acid substitutions follow the residue numbers (cf. amino acid sequences for 3␣-HST and 3␤-HST in Fig. 2A concentrated, and analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting as described previously (9). Protein concentrations were determined by the Bradford method (10).

Enzymatic Assays and Kinetic Analyses
The radiolabeled steroids, [ 3 H]androsterone (54 Ci/mmol) and [ 3 H]pregnenolone (21 Ci/mmol), were purchased from NEN Life Science Products; crystalline steroids were obtained from Steraloids, Inc. (Wilton, NH). 3Ј-Phosphoadenosine 5Ј-phosphosulfate was obtained from Sigma. Sulfotransferase activity was assayed in a buffer consisting of 100 mM Tris-HCl, pH 7.5, and 5 mM magnesium acetate to which 0.1 mM phosphoadenosine 5Ј-phosphosulfate and varying concentrations of radiolabeled steroids (0.05-2.0 M) were added. All reactions were carried out with 1 g of purified protein in a total volume of 100 l for either 2.5 or 5 min at 37°C, unless otherwise specified. Reactions were stopped by the addition of an equal volume of 0.1 N NaOH, and the unreacted steroids were separated from sulfoconjugates by solvent partitioning with 0.8 ml of dichloromethane and vigorous vortexing. The aqueous and organic phases were separated by centrifugation at 13,000 rpm for 3 min. The radioactivity in the aqueous phase was measured by liquid scintillation spectroscopy. Radioactivity in the aqueous phase of unre-acted samples (NaOH addition to enzyme mixture prior to substrate) was used as a blank, and the value was subtracted from each sample.

RESULTS AND DISCUSSION
The aim of these studies was to determine which amino acid residues are responsible for the striking stereospecificity demonstrated by the guinea pig 3␣-HST and 3␤-HST isoforms. These HST isoforms are distinguished by the fact that the ␣-isoform will sulfonate 3-hydroxylated steroids such as androsterone, which have the 3-hydroxyl group spatially oriented in the ␣ position, but will not sulfonate 3-hydroxylated steroids such as pregnenolone, which have the 3-hydroxyl group in a ␤ orientation (Fig. 1A); on the other hand, the ␤-isoform sulfonates 3␤-hydroxylated steroids (pregnenolone), in contrast to 3␣-hydroxylated steroids (androsterone) (Fig. 1B).
Sequence Analysis and Chimera Formation-The 3␣-HST and 3␤-HST isoforms are 87% identical ( Fig. 2A). Although most of the amino acid differences between the two isoforms are scattered throughout the proteins, there are three regions where a substantial number of differences are concentrated ( Fig. 2A). To examine the most NH 2 -terminal of these regions, chimeras were constructed whereby the NH 2 -terminal 91 amino acids of one isoform was joined to the COOH-terminal 196 amino acids of the other isoform. For reasons that are not clear, the chimera consisting of the upstream 91 amino acid region of the ␣-isoform fused to the downstream 196 amino acid region of the ␤-isoform (Fig. 2B) was more highly expressed by a considerable measure than the reverse construct; therefore, the so-called 3␣/␤-HST chimera was employed in these investigations. Interestingly, the 3␣/␤-HST chimera demonstrated a substrate specificity (Fig. 1C) that was nearly identical to the 3␣-HST wild type (cf. Fig. 1A), as distinct from the 3␤-HST wild type (cf. Fig. 1B). This result suggested that the structural determinants regulating the 3-hydroxysteroid substrate stereoselectivity, which characterizes the guinea pig 3-HST isoforms, largely resides within the NH 2 -terminal 91 amino acid region of the proteins. It should be noted that although poorly expressed, the reverse 3␤/␣-HST chimera demonstrated a preference for 3␤-hydroxylated steroids (data not presented).
Amino Acid Substitutions Involving the 3␣/␤-HST Chimera-The ␣ portion of the 3␣/␤-HST chimera contains 12 amino acid residues that are dissimilar from those in the corresponding region of the 3␤-HST isoform (the location of these differences are indicated by arrows in Fig. 2B). Therefore, we set out to examine the relative importance of each amino acid difference in determining the stereoselectivity manifested by the 3␤-HST enzyme. These experiments involved selectively substituting 3␤-HST amino acid residues for corresponding 3␣-HST amino acid residues in the 3␣/␤-HST chimera. Six substituted constructs of the 3␣/␤-HST chimera were generated as presented in Table I. Four of the substituted constructs involving 9 of the 12 amino acid differences continued to exhibit steroid stereospecificity behavior akin to the 3␣-HST wild type and essentially the same as the unaltered 3␣/␤-HST chimera (Figs. 3-6). A fifth substituted construct, however, displayed a clear difference from the previous constructs. That is, a distinction between 3␣-HST and 3␤-HST activity was not evident, and the 3␣-hydroxylated and 3␤-hydroxylated steroid substrates were sulfonated in an essentially equivalent manner (Fig. 7). This result suggested that amino acid residues at position 87 and/or 89 of the 3-HST isoforms might be playing a role in determining substrate specificity. The amino acid residues at position 87 (valine in 3␣-HST and methionine in 3␤-HST) are conservative, suggesting that the amino acids residues at position 89 might be the more critical. The 6th construct, which involved the 12th and final amino acid difference, demonstrated behavior that, for the first time, was more akin to the FIG. 1. Sulfotransferase activity curves for 3␣-HST wild type (A), 3␤-HST wild type (B), and the 3␣/␤-HSTchimera (C). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The HST cDNAs were prepared and overexpressed in bacteria; fusion proteins were columnpurified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures." 3␤-HST wild type and clearly distinct from the unaltered 3␣/ ␤-HST chimera, i.e. it sulfonated the 3␤-hydroxylated steroid substrate more actively than the 3␣-hydroxylated steroid sub-strate (Fig. 8). This last result strongly suggested that the amino acid residue at position 51 plays an especially important role in conferring the characteristic stereospecificity demon-   Table I). The sequence at the top of the figure represents the NH 2 -terminal 91 amino acid ␣-HST portion of the 3␣/␤-HST chimera denoting location of the ␣-HST amino acid residues (arrows) to be replaced by corresponding ␤-HST amino acid residues (bold type above arrows). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated 3␣/␤-HST chimera cDNA was prepared and overexpressed in bacteria; the fusion protein was column-purified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures."  Table I). The sequence at the top of the figure represents the NH 2 -terminal 91 amino acid ␣-HST portion of the 3␣/␤-HST chimera denoting location (arrows) of the ␣-HST amino acid residues to be replaced by corresponding ␤-HST amino acid residues (bold type above arrows). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated 3␣/␤-HST chimera cDNA was prepared and overexpressed in bacteria; the fusion protein was column-purified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures." Table I). The sequence at the top of the figure represents the NH 2 -terminal 91 amino acid ␣-HST portion of the 3␣/␤-HST chimera denoting location (arrows) of the ␣-HST amino acid residues to be replaced by corresponding ␤-HST amino acid residues (bold type above arrows). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated 3␣/␤-HST chimera cDNA was prepared and overexpressed in bacteria; the fusion protein was column-purified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures."

FIG. 5. Sulfotransferase activity curves for mutated 3␣/␤-HST chimera (construct number 3 in
FIG. 6. Sulfotransferase activity curves for mutated 3␣/␤-HST chimera (construct number 4 in Table I). The sequence at the top of the figure represents the NH 2 -terminal 91 amino acid ␣-HST portion of the 3␣/␤-HST chimera denoting location (arrows) of the ␣-HST amino acid residues to be replaced by corresponding ␤-HST amino acid residues (bold type above arrows). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated 3␣/␤-HST chimera cDNA was prepared and overexpressed in bacteria; the fusion protein was column-purified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures."  Table I). The sequence at the top of the figure represents the NH 2 -terminal 91 amino acid ␣-HST portion of the 3␣/␤-HST chimera denoting location (arrows) of the ␣-HST amino acid residues to be replaced by corresponding ␤-HST amino acid residues (bold type above arrows). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated 3␣/␤-HST chimera cDNA was prepared and overexpressed in bacteria; the fusion protein was column-purified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures."  Table I). The sequence at the top of the figure represents the NH 2 -terminal 91 amino acid ␣-HST portion of the 3␣/␤-HST chimera denoting location (arrows) of the ␣-HST amino acid residues to be replaced by corresponding ␤-HST amino acid residues (bold type above arrows). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The HST cDNA was prepared and overexpressed in bacteria; the fusion protein was column-purified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures." strated by the guinea pig 3-HST isoforms.
Amino Acid Substitutions Involving the 3␤-HST Wild Type-To further examine the relative importance of the amino acid residues at position 51 of the guinea pig 3-HST isoforms for determining stereoselectivity, two constructs consisting of single amino acid substitutions in the 3␤-HST wild type were made (Table I). In one construct, asparagine, which is present at residue position 51 in 3␣-HST, was substituted for the isoleucine that normally is present at that position in 3␤-HST. When tested, this construct exhibited activity that was now similar to the first four constructs described above; that is, it displayed behavior more akin to the 3␣-HST wild type and the unaltered 3␣/␤-HST chimera (Fig. 9). As a control, the second construct consisted of a similar substitution, i.e. an asparagine for an isoleucine, made at position 71 of 3␤-HST. When tested, this construct exhibited enzymatic behavior that was indistinguishable from the 3␤-HST wild type enzyme (Fig. 10). The results of these latter experiments strengthen the case for the amino acid residue at position 51 as playing a crucial role in determining the 3-hydroxy stereospecificity that characterizes the guinea pig 3␣-HST and 3␤-HST isoforms.
Summary- Table III summarizes the kinetic values determined for the 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, the substrates used to evaluate the stereospecificity of each described guinea pig HST preparation. It is concluded, based on the data presented, that the structural determinants underlying the stereoselectivity exhibited by the guinea pig 3␣-HST and 3␤-HST isoforms largely resides within the NH 2 -terminal 91 amino acid region of the proteins. Specifically, the amino acid residue at position 51 is crucial; if asparagine is present at that position, 3␣-HST activity prevails, whereas if isoleucine is present, 3␤-HST activity is dominant. In addition, the amino acid residues at position 87 and/or 89 might also play a significant role, although that possibility has not been tested. Finally, although these recombinant structure-function studies have yielded interesting and provocative results, a more precise understanding of the structural determinants involved in imparting the exceptional stereospecificity that is characteristic of the guinea pig 3-HST enzymes must await crystallization and final resolution of their three-dimensional structures.
The only steroid sulfotransferase whose structure has been solved is mouse estrogen sulfotransferase; however, it is only 31% identical to the guinea pig 3␣-HST and 3␤-HST enzymes (11). Furthermore, the amino acid residues, particularly those within the NH 2 -terminal portion of mouse estrogen sulfotransferase that are found near the estradiol-17␤ substrate, although conserved in guinea pig estrogen sulfotransferase (12),  Table I). The 3␣-hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated ␤-HST cDNA was prepared and overexpressed in bacteria; the fusion protein was columnpurified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures."  Table I). The 3␣hydroxylated steroid, androsterone, and the 3␤-hydroxylated steroid, pregnenolone, substrates are indicated. The mutated ␤-HST cDNA was prepared and overexpressed in bacteria; the fusion protein was columnpurified, cleaved, and assayed for sulfotransferase activity as described under "Experimental Procedures."

TABLE III
Kinetic values for androsterone and pregnenolone substrates Wild type guinea pig 3␣-HST and 3␤-HST, 3␣/␤-HST chimera, site-selected mutants of 3␣/␤-HST chimera, and 3␤-HST wild type were overexpressed in Escherichia coli, purified, and assayed for sulfotransferase activity as described under "Experimental Procedures." Values represent the average of triplicate determinations. a Construct: ␣WT, ␤WT, and ␣/␤ refer to 3␣-HST wild type, 3␤-HST wild type, and 3␣/␤-HST chimera, respectively; numbers 1-8 correspond to the mutant constructs listed in Table I. are not conserved in the guinea pig 3-HST isoforms. The amino acid residue in mouse estrogen sulfotransferase that would appear comparable with residue 51 in the guinea pig 3-HST isoforms is a serine, and this serine is not considered to be involved in the binding of the estradiol-17␤ substrate (11). Based on analysis of the mouse estrogen sulfotransferase structure, the most highly conserved region of the molecule is involved in interaction with the sulfonate donor compound, phosphoadenosine 5Ј-phosphosulfate, whereas the region interacting with substrate appears to be much more variable (11).