(Received for publication, January 24, 1997, and in revised form, June 18, 1997)
From the Ohio State Biochemistry Program and
§ Department of Veterinary Biosciences, Ohio State
University, Columbus, Ohio 43210
Human growth hormone (hGH) stimulates somatogenic and lactogenic actions through the GH and prolactin (PRL) receptors, respectively. In contrast, non-primate GHs stimulate only somatotropic action. Phe44, of the human GH sequence is present in all hormones stimulating lactogenic action and absent in all hormones stimulating only somatotropic action. We speculate that the presence of Phe44 is a feature necessary for specifying lactogenic activity. In this report, the role of Phe44 was investigated by its deletion or substitution with alanine or leucine. Deletion of Phe44 or substitution with leucine did not significantly change the structure of hGH as determined by circular dichroism, absorbance, and fluorescence spectroscopies. In contrast, substitution of alanine perturbed the structure. Deletion of Phe44 reduced binding affinity for the lactogenic receptor, resulting in a reduced activation. Substitution with either alanine or leucine partially restored lactogenic receptor binding affinity, which correlated with the hormones' activity in the Nb2 rat lymphoma cells. All the recombinant hGHs had similar somatotropic activities in FDC-P1 cells transfected with the hGH receptor. These data indicate that the hydrophobic side chain of Phe44 is required for lactogenic receptor binding and activation but is unnecessary for somatotropic action.
Growth hormones (GH)1 are members of the GH/prolactin (PRL) family which promote a diverse group of biochemical and physiological processes (1) through the GH receptor (2). The GH receptor is a member of a subclass of the hematopoietic receptor superfamily which includes the receptors for cytokines; PRL; interlukins 2, 3, 4, 6, and 7; and erythropoietin (3). Hematopoietic receptors each have a single transmembrane sequence with structural similarities in their extracellular domains. Generally, receptors in this superfamily are activated by ligand-induced homo- or heterodimerization resulting in activation of the JAK family of kinases (2, 4). GH activates its receptor by homodimerization of two receptor molecules; each receptor binds distinct sites on the hormone (5).
Primate GHs are unique to the GH family because they bind and activate PRL receptors as well as the GH receptors. PRLs are not somatotropic (6) while non-primate GHs are not lactogenic. Human GH (hGH) is a well characterized molecule to study structure-function relationships because of the availability of multiple crystal structures (7-9), extensive information from mutagenic studies (10-13), as well as the ability to activate lactogenic and somatotropic receptors.
The three-dimensional structures of hGH bound to the hGH receptor in a 1:2 complex (7) or the hPRL receptor in a 1:1 complex (8) have been reported. Association with either receptor at site 1 buries similar surfaces of hGH while the overall three-dimensional structures of hGH extracted from these complexes can be nearly superimposed (8), suggesting that association with either receptor does not induce large changes in conformation. Alanine scanning of hGH residues that contact either the hGH or hPRL receptor has identified two overlapping sets of residues critical for providing the free energy necessary to facilitate binding. These data sets have been used to engineer variants of hPRL (11) and human placental lactogen (14) to bind the hGH receptor at site 1 with affinities close to that of wild-type hGH. These studies have stressed the energetically important contributions necessary for binding either receptor at site 1. Yet, other residues that are not energetically important may be structurally important in site 1 binding to either somatotropic or lactogenic receptors.
Phe44 of hGH is one residue which may be of structural importance in site 1 binding to somatotropic or lactogenic receptors. Phe44 is located in the middle of mini-helix 1 and is involved in a hydrophobic interaction with leucine 157 and tyrosines 160 and 164 of the amino terminus of helix 4. This mini-helix is buried at the site 1 receptor interface upon binding to either receptor with a significant 2.5 Å movement toward the amino terminus of helix 4 when hGH is bound to the lactogenic receptor rather than the somatotropic receptor (3). Amino acid sequence comparisons of the PRLs and GHs found Phe44 to be present in all lactogenic hormones but absent in all non-primate GHs, suggesting that Phe44 may be required for lactogenic action even though it does not contact either receptor.
To examine the requirement of Phe44 for lactogenic and somatotropic actions in hGH, we have deleted it. We hypothesize this deletion will mimic non-primate GHs, characterized by somatotropic activity without lactogenic activity. The deletion of Phe44 may affect the actions of hGH by altering the relative helical positions of the residues of mini-helix 1 that contact the receptor or by removal of a necessary hydrophobic side chain. Therefore, the positional effects were studied by replacement of Phe44 with alanine or the more hydrophobic leucine.
An f1
origin of replication was inserted at a unique ClaI site in
pT7-7 (kindly provided by S. Tabor, Harvard Medial School, Boston, MA).
The negative strand pT7-7 phagemid was used for cloning, production of
single-stranded DNA, and expression of hGH. Escherichia coli
strains DH5, RZ1032 (dut
, ung
), and
BL21(DE3) were used for cloning, uridine-substituted single-stranded DNA production, and protein expression, respectively.
Poly(A) mRNA was isolated from
a human pituitary (Cooperative Human Tissue Network) and cDNA was
prepared. A polymerase chain reaction amplified the coding sequence for
mature hGH using the following 5 and 3
primers: 5
primer,
5
-CCCCGAATTCGAAGGAGATATACATATGTTCCCGACTATCCCGCTTTCCAGGCCTTTTGACAACC-3
; 3
primer, 5
-CCCCAAGCTTGGCTAGAAGCCACAGCTGCC-3
.
The 5
primer was designed to introduce a methionine codon
(underlined) at the start of the mature sequence, change the first
seven codons to those preferred by bacteria, and add a unique
NdeI (italics) site encompassing the initiation codon. The
3
primer introduced a unique HindIII (italics) site
following the termination codon (underlined). Amplification products
were purified, digested with NdeI and HindIII and
the 587-base pair fragment ligated into the pT7-7f(
) phagemid.
Products were transformed into DH5
cells, selected by resistance,
screened by restriction digests, and confirmed by DNA sequencing
(15).
In vitro mutagenesis
was performed by the Kunkel method (16, 17). Primers were designed to
delete Phe44 (Phe44) or replace
Phe44 with alanine (F44A) or leucine (F44L). Selection for
positive mutants was facilitated by removal of an adjacent
PstI site through a primer-introduced translationally silent
mutation. Clones were evaluated by restriction digest with
PstI. The complete hGH DNA sequence in clones uncut by
PstI was determined to confirm the presence of the desired
mutation.
Purified phagemids
containing wild-type, Phe44, F44A, or F44L hGH DNAs were
transformed into BL21(DE3) cells and transfected clones selected by
antibiotic resistance and confirmed by restriction studies. One-liter
cultures were grown to an A600 of 0.3 and
induced with 0.4 mM
isopropyl-
-D-thiogalactopyranoside. Incubation was continued for 4 h. Cells were harvested by centrifugation and lysed by two passes through a French pressure cell in 50 ml of 100 mM Tris, pH 7.5, 25 mM dithiotheritol, and 1 mM phenylmethylsulfonyl fluoride. Inclusion bodies were
collected by centrifugation, resuspended in 100 ml of 4.5 M
urea, 50 mM Tris, and the pH was adjusted to values between
11.0 and 11.5. Inclusion bodies were solubilized for 2 h at room
temperature. Solubilized inclusion bodies were clarified by
centrifugation and air-oxidized with stirring at 4 °C for 1-2 days.
Oxidized protein was dialyzed into 20 mM Tris, pH 7.4 (four
times, 4 liters each). Just prior to purification the pH was adjusted
to 9.0 and hGH was purified using DE52 anion exchange resin (Whatman,
Clifton, NJ). The protein was eluted using a 0-750 mM NaCl
gradient in 20 mM Tris, pH 9.0. Eluted protein was dialyzed
into 5 mM ammonium bicarbonate (six times, 4 liters each)
and lyophilized.
Proteins were
evaluated for size and purity by 12% SDS-polyacrylamide gel
electrophoresis under either nonreducing or reducing conditions.
Protein concentrations were calculated using 277 nm of
a 0.1% solution = 0.82 cm
1 (18). Absorption,
fluorescence, and far-UV circular dichroism spectra were collected at
20 °C in 10 mM Tris, pH 8.2, 150 mM NaCl.
Nb2 rat lymphoma cells (19) were maintained and manipulated as described previously (16). Recombinant hGHs stock concentrations were determined by the bicinchoninic acid/copper sulfate assay (20). Stocks were diluted in assay media and added to triplicate wells (96-well plate, Costar, Cambridge, MA) each containing 20,000 Nb2 cells in a final volume of 100 µl with various hGH concentrations. Cells were incubated at 37 °C in a 5% CO2, 95% air atmosphere for 48 h. Proliferation was assessed by a vital dye method with the addition of 10 µl of Alamar Blue (Accumed International, West Lake, OH) per well and continued incubation for 4 h. The oxidation/reduction of Alamar Blue was then evaluated at 570 and 600 nm using a microplate reader (Molecular Devices, Palo Alto, CA). Absorbances at 570 and 600 nm were used to calculate the percent reduction of the dye that is correlated with cell number (r2 > 0.99). The percent reduction was then used to calculate an ED50 by a four-parameter fit method (21).
FDC-P1 Somatotropic AssayFDC-P1 cells containing the hGH receptor were a gift from Genentech (South San Francisco, CA). Cells were maintained in RPMI 1640 containing 10 µM 2-mercaptoethanol, 10 units of interleukin 3/ml, and 10% fetal calf serum (22). Log phase cells were collected and washed three times with RPMI 1640. Washed cells were resuspended in media devoid of interleukin 3 and returned to the incubator for 24 h prior to the assay. Recombinant hGH stock concentrations were determined as above. Stocks were diluted with interleukin 3-free medium to the desired concentration and added to triplicate wells with 15,000 FDC-P1 cells in a total volume of 100 µl with various hGH concentrations. Plates were gently vortexed and then incubated at 37 °C in a 5% CO2, 95% air atmosphere for 48 h. Proliferation was assessed as described above. The percent reduction of Alamar Blue was used to calculate an ED50 by a four-parameter fit method (21).
Binding AssayshGH was iodinated with Iodogen (Pierce) and carrier-free [125I]iodine to a specific activity of 21 µCi/µg. Binding reactions contained membranes from 2 × 106 cells in 700 µl of Fisher's medium supplemented with 0.5% bovine albumin, 5 mM MgCl2, 1 mM ZnSO4, 1 mM phenylmethylsulfonyl fluoride, 10 µg/mL aprotinin, approximately 1.9 ng of 125I-hGH and various concentration of recombinant hGHs. Incubations were run for approximately 20 h at room temperature. Membranes were collected by centrifugation and associated 125I-hGH measured. Data from competition studies were used to calculate relative affinities by the method of Scatchard (23).
Phagemids contained the mature DNA sequence for wild-type hGH with the addition of an initiation methionine. Dideoxy sequencing of the complete wild-type and mutant inserts confirmed the wild-type code (24) and the desired mutations.
Wild-type and mutant hGHs were expressed and purified with yields of
approximately 40 mg/liter of fermentation. The proteins co-migrated
with National Hormone and Pituitary Program (NIH) hGH on SDS
polyacrylamide gel electrophoresis under reducing conditions and were
observed to be greater than 95% homologous (Fig.
1). Co-migration of recombinant proteins
with NIH hGH on nonreducing SDS-polyacrylamide gel electrophoresis gels
suggested that all proteins were folded to produce correct structures.
A small portion of several preparations formed dimers that were evident
on nonreducing SDS-polyacrylamide gel electrophoresis gels.
Spectroscopy of Recombinant Proteins
The circular dichroism
spectra of wild-type and mutant hGHs were very similar in shape (Fig.
2). Phe44 and F44L hGHs
had molar ellipticity values at 222 nm of 93 and 100% when compared
with wild-type hGH, suggesting insignificant differences resulting from
these mutations. In contrast, substitution of alanine at position 44 induced a 50% reduction of the 222-nm signal.
Absorbance spectra (Fig. 3) of the
wild-type and Phe44 hGHs overlay each other. Replacement
of Phe44 with leucine or alanine increased absorption at
277 nm (see raw data inset in Fig. 3). When the data are
normalized for absorption at 277 nm, F44L and F44A have an increased
absorption in the 250-260-nm region.
The shape and maxima of the fluorescence spectra (excitation 285 nm)
(Fig. 4) for the wild-type,
Phe44, and F44L hGHs were similar. In contrast, F44A hGH
had a reduced quantum yield (Fig. 4, inset).
Biological Activity of the Recombinant Proteins
In the FDC-P1
somatotropic assay (Fig. 5, Table
I) the wild-type, Phe44,
F44L, and F44A hGHs had ED50 values that varied over a
range of 2.9-fold between 187 and 531 pM. The assays had an
average coefficient of variation of 5.6%. Wild-type and F44L hGH
appeared to be indistinguishable from each other, while
Phe44 and F44A hGH had similar activities. When the
dose-response curves for these assays are compared, the maximal
cellular growth responses were similar. A pituitary isolate of hGH
(NIH) had an ED50 of 78 pM. The F44A hGH was
the least potent preparation with an increased ED50 of 531 pM.
|
The lactogenic activity of the recombinant hGHs was assessed in the Nb2
biological assay (Fig. 6, Table I) where
the average coefficient of variation was 5.0%. Wild-type hGH had an
ED50 of 8.0 pM, compared with a 5.5 pM ED50 for hGH isolated from the pituitary
(NIH). Elimination of Phe44 severely reduced the biological
activity of hGH increasing the ED50 218-fold to 1,746 pM. Replacement of Phe44 with alanine increased
the ED50 to 546 pM, while replacement with the
more hydrophobic leucine gave an ED50 of 158 pM. Addition of sufficient hormone showed a equivalent
maximal response for each mutant when compared with wild-type hGH.
Binding Activity of the Recombinant Proteins
The relative abilities of wild-type and mutant hGHs to bind the lactogenic (Nb2 cells containing the short form of the receptor) or somatotropic (FDC-P1 cells transfected with the hGH receptor) receptors were measured. The affinity of wild-type hGH was 0.43 nM and 0.61 nM for lactogenic and somatotropic receptors, respectively (Table I). Deletion of Phe44 reduced the relative binding to the lactogenic receptor by 126-fold, while reducing the affinity for the somatotropic receptor by only 2.5-fold. As in bioassays, replacement of Phe44 with alanine or the more hydrophobic leucine increased the abilities of these mutants to bind the lactogenic receptor. These mutations had minimal effects in the somatotropic receptor assays. When the loss of lactogenic action relative to somatotropic action for receptor binding affinity and activity (Table I) were correlated for each hormone, the correlation between binding and activity was high (r2 = 0.765), suggesting an association between binding and activity.
Deletion of Phe44 in methionyl-hGH reduced lactogenic activity and receptor binding affinity but retained somatotropic activity and receptor binding affinity when compared with the wild-type hGH (Table I). Replacement of Phe44 with alanine or leucine provides increased binding to the lactogenic receptor and lactogenic action with increasing bulk and hydrophobic character at position 44. These studies demonstrate that Phe44 is a necessary structural feature for lactogenic action, even though Phe44 does not contact the either receptor when bound at site 1.
Deletion of Phe44 eliminates the protein backbone in this
position and rotates the adjacent portions of mini-helix 1 approximately 100° relative to each other. Replacement of
Phe44 with alanine or leucine retains the periodic
relationships of residues within mini-helix 1 while providing side
chains with differing degrees of hydrophobic character. These
replacements produced hormones with high somatotropic activity but
lactogenic activities intermediate between Phe44 and the
wild-type hGH. These mutants demonstrate a correlation between
increasing hydrophobic character at position 44 and the ability to
induce lactogenic action. These results suggest the important
structural feature at position 44 is the presence of a large
hydrophobic residue and not the relative orientation of residues in
mini-helix 1.
Phe44 contributes to a hydrophobic pocket composed of
leucine 157 and tyrosines 160 and 164 located at the amino terminus of helix 4 (25). hGH assumes two slightly different conformations of this
hydrophobic cluster when bound to either the somatotropic or lactogenic
receptor (Fig. 7). Mini-helix 1 is more
closely packed in the hydrophobic cluster when bound to the lactogenic receptor. We speculate that elimination of Phe44 disrupts
the hydrophobic packing prohibiting the lactogenic conformation which
manifests as a reduction of lactogenic receptor binding affinity. The
reduced receptor binding affinity is reflected by an increased
ED50 for biological activity.
Cunningham and Wells (13) have shown that deletion of residues 32-46, inclusive of Mini-helix 1, reduced the affinity for the lactogenic receptor by 6.1-fold. This modest change in affinity suggests that these residues are not crucial for lactogenic receptor binding. In contrast, our data indicate that the elimination of Phe44 decreases the affinity for the lactogenic receptor by 126-fold (Table I). While elimination of mini-helix 1 obviates the need for hydrophobic packing of mini-helix 1; the presence of mini-helix 1 requires appropriate hydrophobic packing. Phe44 appears to be critical for packing hGH in a conformation compatible with lactogenic receptor binding.
Comparison of the somatotropic bioassay data indicates that all mutant hGHs have similar biological activities. This is interesting considering the significant changes observed in the circular dichroism, absorbance and fluorescence spectra for F44A hGH from that of the other proteins. Changes in fluorescence and UV absorption spectra suggest increased hydration of the aromatic residues but do not indicate a significant perturbation of the overall structure. In contrast, the circular dichroism spectrum of F44A hGH suggests a 50% loss in helical content. However, the biological assay data indicates that the overall structure is either intact or is able to be recovered during receptor binding as demonstrated by the minimal 2.9-fold increase in the somatotropic ED50 produced by alanine substitution. The significant decrease observed in the 222 nm molar ellipticity of F44A hGH is consistent with the reduction reported when wild-type bGH is reduced (26). However, detection of free sulfhydryl groups using Ellman's reagent (27) indicated that less than 5.0% of all disulfide bridges were broken in the F44A mutant, a value similar to that observed with wild-type hGH. At this time, no plausible explanation can be offered for the altered spectroscopic properties of F44A hGH.
In conclusion, deletion of Phe44 from mini-helix 1 was spectroscopically indistinguishable from wild-type hGH. This deletion essentially eliminated lactogenic action, while minimally affecting somatotropic action. Further, the ability of the deletion mutant to stimulate lactogenic action could be recovered by introduction of increasingly hydrophobic residues. Mini-helix 1 of hGH is irrelevant to lactogenic or somatotropic receptor binding at site 1, but the interaction with the first receptor functions to position mini helix 1 (containing Phe44) to induce subtle changes in the protein scaffolding of the hormone. We speculate these receptor-induced structural changes display the dimeric complex in an orientation that allows binding by a second homologous receptor.