Sequence polymorphisms at the growth hormone GH1/GH2-Nand GH2-Z gene copies and their relationship with dairy traitsin domestic sheep (Ovis aries)

G. M. Vacca • M. L. Dettori • F. Balia •

S. Luridiana • M. C. Mura • V. Carcangiu •

M. Pazzola

Received: 25 July 2012 / Accepted: 30 April 2013 / Published online: 8 May 2013

� Springer Science+Business Media Dordrecht 2013

Abstract The purpose was to analyze the growth hor-

mone GH1/GH2-N and GH2-Z gene copies and to assess

their possible association with milk traits in Sarda sheep.

Two hundred multiparous lactating ewes were monitored.

The two gene copies were amplified separately and each

was used as template for a nested PCR, to investigate

single strand conformation polymorphism (SSCP) of the

50UTR, exon-1, exon-5 and 30UTR DNA regions. SSCP

analysis revealed marked differences in the number of

polymorphic patterns between the two genes. Sequencing

revealed five nucleotide changes at the GH1/GH2-N gene.

Five nucleotide changes occurred at the GH2-Z gene: one

was located in exon-5 (c.556G [ A) and resulted in a

putative amino acid substitution G186S. All the nucleotide

changes were copy-specific, except c.*30delT, which was

common to both GH1/GH2-N and GH2-Z. Variability in

the promoter regions of each gene might have conse-

quences on the expression level, due to the involvement in

potential transcription factor binding sites. Both gene

copies influenced milk yield. A correlation with milk

protein and casein content was also evidenced. These

results may have implications that make them useful for

future breeding strategies in dairy sheep breeding.

Keywords Domestic sheep � oGH gene � Polymorphism �GH2-N � GH2-Z � Sheep growth hormone

Introduction

The Growth Hormone (GH) is a peptide hormone 191

amino acid long, synthesized by the anterior pituitary

gland. The growth hormone has two main actions closely

related: it stimulates body growth and regulates cell

metabolism, as it is involved in the metabolism of proteins,

carbohydrates, lipids and minerals. It is demonstrated that

this molecule plays an important role in increasing growth

performance and milk yield in domestic animals [1]. For

this reason the GH gene is considered a candidate marker

for productive traits in livestock. Several studies reported

the effects of the GH gene polymorphism on productive

traits in dairy cattle [2, 3], goats [4, 5] and sheep [6]. In

sheep, the GH gene (oGH), is mapped to chromosome 11

(11q25) [7], it is about 1.8 kb long and contains five exons.

Two alleles have been found at this locus, named Gh1 and

Gh2, the GH1 gene copy occurs at the Gh1 allele, and the

GH2-N and GH2-Z gene copies occur at the Gh2 allele,

which is duplicated [8]. The Gh2 allele occurs in about

90 % of the populations studied so far, suggesting that it

may have a selective advantage [6]. Based on sequence

analysis of ewes homozygous for the Gh1 and Gh2 alleles,

Ofir and Gootwine [7] revealed that within the Gh1 allele,

the sequence of the GH1 gene is highly conserved, while

sequence differences have been evidenced between the

GH2-N and GH2-Z copies. The GH2-N gene is expressed

in the pituitary [9] and GH2-Z is expressed in the placenta

[10].

Sequencing of the 3.5 kbp long intercopy region,

extending across the GH2-N and GH2-Z copies, has

revealed the occurrence of a DNA region, localized at the

50UTR of GH2-Z, which differentiates the GH1/GH2-N

copies from GH2-Z, allowing to set up a selective PCR

amplification [6].

G. M. Vacca � M. L. Dettori (&) � F. Balia � S. Luridiana �M. C. Mura � V. Carcangiu � M. Pazzola

Dipartimento di Medicina Veterinaria, Universita degli Studi

di Sassari, via Vienna 2, 07100 Sassari, Italy

e-mail: mldettori@uniss.it

123

Mol Biol Rep (2013) 40:5285–5294

DOI 10.1007/s11033-013-2629-9

The purpose of this study was to analyze, by Single

Stranded Conformation Polymorphism (SSCP) and

sequencing, genetic variability of the 50UTR (Untranslated

Region), exon-1, exon-5 and 30UTR, at the GH1/GH2-N and

GH2-Z gene copies and to assess their possible association

with milk production and composition in Sarda sheep.

Materials and methods

Animals and samples

Two hundred lactating Sarda breed ewes were randomly

selected from four farms, with similar traditional man-

agement and feeding system, located in Sardinia (Italy), in

an area between 40 degrees North latitude and 8 degrees

Est longitude. The sheep were healthy and in good state of

nutrition, multiparous, aged between 4 and 5 years, in their

third or fourth lactation. Once a month, from each sheep,

from February to June, daily milk yield was recorded and a

milk sample was collected. Individual blood samples were

taken for DNA extraction.

The Sarda sheep breed is the Italian most numerous

dairy breed with about 3.3 million heads reared only in

Sardinia [11] where 5 % of the world’s sheep milk is

produced [12]; for its characteristics of rusticity and high

milk yield, this breed is reared also in other Italian regions

and in some countries of the Mediterranean area.

Milk analysis

Milk samples were analyzed by using an infrared spec-

trophotometer (Milko-Scan 133B; Foss Electric, DK-3400

Hillerød, Denmark) to assess fat and protein percentage

according to the International Dairy Federation standard

(IDF 141C:2000) and casein content (FIL-IDF 29:1964).

DNA analysis

DNA extraction was performed with a commercial kit

(NucleoSpin Blood, Macherey–Nagel) and the DNA con-

centration and purity were measured by spectrophotometer

(Eppendorf Biophotometer, Hamburg, Germany). The

GH1/GH2-N gene was selectively amplified utilizing the

primer pair GHTF/GHTR [6] to obtain a DNA fragment of

about 2.05 kbp. In order to selectively amplify the GH2-Z

allele, a primer pair was designed with the Primer3 soft-

ware (http://frodo.wi.mit.edu/primer3/), based on GenBank

Acc. No. DQ461643, to get a fragment of about 2.22 kbp

(Table 1). The PCR reaction was carried out in a final

volume of 25 ll, with 25–50 ng of genomic DNA;

4–16 pmol of each primer; 1,5 mM of MgCl2; 200 lM of

dNTPs, 1 U of Taq Platinum DNA Polymerase (Invitrogen,

Life Technologies). The PCR program consisted of a

denaturing step at 94 �C for 5 min, followed by 35 cycles

[94 �C (30 s), 58 �C (20 s), 72 �C (2 min)] and an elon-

gation step at 72 �C for 5 min.

The GH1/GH2-N and GH2-Z genes were used sepa-

rately as template for a nested PCR, aimed at the amplifi-

cation of the 50UTR (fragment I, 125 bp), exon-1 (fragment

II, 112 bp), exon-5 (fragment VI, 289 bp) and 30UTR

(fragment VII, 150 bp), using the primer pairs proposed by

Marques et al. [6]; while the primer pair GH2ZF/GH5PR

was used for the nested PCR of the 50UTR of the GH2-Z

gene (fragment I, 358 bp) (Fig. 1). The two DNA frag-

ments I and II were partially overlapping, having in com-

mon about 60 bp, and the DNA segments VI and VII had in

common about 87 bp.

The nested PCR products were analyzed by SSCP.

Analysis was carried out on a D-Code Universal Mutation

Detection System (BioRad), as follows: 4 ll of each PCR

product was dissolved in 15 ll of denaturation solution

(95 % of formamide, 10 mM NaOH, 0.05 % of xylene-

cyanol and 0.05 % of bromophenol blue), heat-denaturated

at 95 �C for 5 min, chilled on ice and loaded onto 10 %

polyacrylamide gels (acr/bis 29:1) with TBE 0.59. Run-

ning conditions: 12 �C constant temperature, except frag-

ment I of GH1/GH2-N which ran at 4 �C; 25 W constant

power and 4–6 h running time. Products were visualized

with Sybr-Gold Nucleic Acid Gel Stain (Invitrogen, Life

Technologies) for 30 min; the different electrophoretic

profiles were displayed using a UV-transilluminator

(UVITEC, Cambridge). From one to three DNA fragments,

showing the same SSCP profile, were sequenced in both the

forward and reverse directions, using the primer pairs

GHTF/GH1R and GH5F/GHTR for the GH1/GH2-N gene

and the primer pairs GH2ZF/GH1R and GH5F/GH2ZR for

the GH2-Z gene. Sequencing was achieved with an Applied

Biosystems 3730 DNA Analyzer (Applied Biosystems,

Foster City, CA, USA), after purification with Agencourt�

AMPure� kit (Beckman Coulter, USA).

The Finch TV software (http://www.geospiza.com/

Products/finchtv.html) was used to view and edit chro-

matograms, and the BioEdit software [13] was used to

align sequences. The nucleotide variations were described

according to the Human Genome Variation Society (http://

www.hgvs.org/mutnomen/). Allele frequencies and Hardy–

Weinberg equilibrium were determined with POPGENE

V1.32 software [14]. Alibaba 2.1 software program, with

the TRANSFAC database version 7.0 (http://www.gene-

regulation.com/pub/programs/alibaba2) were used to ana-

lyse the promoter region, for putative transcription factor

binding sites. Haplotypes were predicted, based on linkage

disequilibrium and allele frequencies, within each gene

copy (GH1/GH2-N and GH2-Z) using the PHASE program

[15].

5286 Mol Biol Rep (2013) 40:5285–5294

123

Statistical analysis

Association analysis between the GH1/GH2-N and GH2-Z

genotypes and milk traits with repeated individual obser-

vations was performed using a mixed model (SAS Inst.

Inc., Cary, NC) which considered stage of lactation (days

in milk, DIM), genotype and the interaction between stage

of lactation and genotype as fixed effects. Animal nested

within genotype was considered as random effect. For all

parameters, model effects were declared significant at

P \ 0.05. Multiple comparisons of the means were per-

formed using the Bonferroni’s method [16]. Only geno-

types present in at least 2 % of the population were

considered.

Results

The GH1/GH2-N and GH2-Z copies of the ovine growth

hormone gene were amplified separately, to obtain two

DNA fragments about 2,055 and 2,219 bp long,

respectively. PCR amplification of the GH2-Z allele was

successful for all 200 ewes analysed, indicating that no

homozygous Gh1/Gh1 genotypes occurred in the present

study. However, it cannot be excluded that heterozygous

Gh1/Gh2 subjects were present. Each gene copy was used

as template for a nested PCR in order to investigate SSCP

polymorphism of 50UTR and exon-1 (fragments I and II);

exon-5 and 30UTR (VI and VII).

The images of the GH1/GH2-N gene SSCP are shown in

Fig. 2 and SSCP results are summarized in Table 2. The

GH1/GH2-N fragment I revealed three different patterns,

the most frequent was pattern A, and sequencing revealed

that patterns A and C had homozygous genotype combi-

nations, while pattern B (more rare in the population ana-

lyzed) was heterozygous. Fragment II gave four patterns,

the most frequent was pattern A, and sequencing revealed

that patterns A and D were homozygous, while patterns B

and C were heterozygous. Single strand polymorphism of

fragments VI and VII was characterized for the presence of

five different patterns, showing equal frequencies, as the

mutations detected were all located in the overlapping

Table 1 Primer sequences for

analysis of the sheep GH geneFragment name Primer name Primer sequence (50–30) Reference

GH1/GH2-N GHTF CCAGAGAAGGAACGGGAACAGGATGAG Marques et al. [6]

GHTR ATAGAGCCCACAGCACCCCTGCTATTG

GH2-Z GH2ZF TGGCTACACCTCTTCCTGCT This paper

GH2ZR GGAGGAACCGGGTCAATTAT

I GH5PF GGGAAAGGGAGAGAGAAGAAGCCAG Marques et al. [6]

GH5PR CAGCCATCATAGCTGGTGAGCTGTC

II GH1F CAGAGACCAATTCCAGGATC Marques et al. [6]

GH1R TAATGGAGGGGATTTTCGTG

VI GH5F CCCTTGGCAGGAGCTGGAAG Marques et al. [6]

GH5R AAAGGACAGTGGGCACTGGA

VII GH3PF CCTTCTAGTTGCCAGCCATCTGTTG Marques et al. [6]

GH3PR CCACCCCCTAGAATAGAATGACACCTAC

Fig. 1 Schematic representation of sheep GH gene. Exons are

represented by black boxes. The GH1 gene belongs to the Gh1

allele, while genes GH2-N and GH2-Z belong to the Gh2 allele. The

GH1/GH2-N genes have same sequences in the flanking regions. The

GH2-Z gene was amplified selectively thanks to the sequence stretch

of the intercopy region that differs from the corresponding regions in

the GH1/GH2-N genes. The DNA fragments I, II, VI and VII,

analysed by nested PCR, are framed

Mol Biol Rep (2013) 40:5285–5294 5287

123

region. Sequencing revealed that patterns A and D showed

homozygous genotype combinations.

Images resulting from SSCP analysis of the GH2-Z gene

are shown in Fig. 3 and results from polymorphism anal-

ysis are displayed in Table 2. Fragment I revealed three

different patterns. Sequencing revealed that pattern B,

which was the most frequent, showed a heterozygous

genotype combination, in contrast, fragment II was

monomorphic. Fragment VI was characterized for the

presence of five different patterns. Sequencing revealed

that only patterns A (the most frequent) and E (the least

frequent) had homozygous genotype combinations. Frag-

ment VII showed three different patterns, only pattern B

was heterozygous.

Sequencing of the SSCP polymorphic samples revealed

the nucleotide (nt) variations described in Table 3. At the

GH1/GH2-N gene, five nt changes were identified, three

were located in the 50UTR and two in the 30UTR, compared

to acc. no. DQ450146. Both exon-1 and exon-5 were

monomorphic. The nt variations detected with respect to

the reference sequence had low frequencies, except for the

deletion c.*30delT, which had a frequency of 0.843. All the

nt changes were in HW equilibrium for the GH1/GH2-N

locus, except c.-19T [ A and c.-15G [ A.

Five nucleotide changes occurred at the GH2-Z gene;

compared to acc. no. DQ46164. Two were located at the

50UTR, two at the 30UTR; one was located in exon-5

(c.556G [ A) and resulted in a putative amino acid substi-

tution G186S. Exon-1 was monomorphic. Allele frequencies

revealed that c.556G [ A was rare, and it was the only nt

variation out of the HW equilibrium, for this locus.

Haplotype prediction at the GH1/GH2-N locus revealed

a total of 7 haplotypes with frequencies [0.01, and only

two showed frequencies[0.05: haplotypes 1 (83 %) and 2

(6.3 %) (Table 4). At the GH2-Z locus, 6 haplotype com-

binations with frequency [0.01 were estimated, three of

them had frequency [0.05: haplotypes 1 (60 %), 2

(23.2 %) and 3 (12.5 %) (Table 4).

The effect of genotype and stage of lactation at GH1/

GH2-N and GH2-Z genes on milk yield, fat, protein and

casein percentage are shown in Tables 5 and 6. Statistical

analysis was performed considering only DNA fragments

II and VI for GH1/GH2-N gene and fragments I and VI for

the GH2-Z gene, as they were the most informative. The

stage of lactation significantly influenced (P \ 0.001) the

milk production traits considered here, as shown in Fig. 4.

A significant interaction effect between GH1/GH2-N

fragment VI genotype x DIM was evidenced on milk yield

and between GH2-Z fragment I genotype x DIM on fat

percentage (Table 5). Statistical analysis revealed that both

copies of the sheep GH gene affected phenotypic variation

of milk traits. Ewes displaying pattern A of fragment II and

pattern C of fragment VI (GH1/GH2-N) produced the

highest (P \ 0.05) milk yield, and pattern A of fragment

VI was significantly (P \ 0.05) associated with the highest

protein and casein percentage (Table 6). Ewes with pattern

A of fragment I (GH2-Z) showed the highest (P \ 0.05)

milk yield, and patterns B and D of fragment VI were

Fig. 2 SSCP patterns of GH1/

GH2-N gene DNA fragments

I (50UTR), II (exon-1), VI

(exon-5) and VII (30UTR). A

single letter designation was

adopted to indicate the different

patterns. Each pattern differs

from the others depending on

the relative position of the

single strand DNA bands on

polyacrylamide gel. The

genotype combination

corresponding to each pattern is

indicated in Table 2

5288 Mol Biol Rep (2013) 40:5285–5294

123

Table 2 SSCP pattern frequencies and genotype combinations found at the GH1/GH2-N and GH2-Z genes of Sarda sheep

DNA

fragment

GH1/GH2-N GH2-Z

Pattern frequencies

(%)

Genotype

combinations

Nta Pattern frequencies

(%)

Genotype

combinations

Ntb

I A (94.5) TT/GG 976/980 A (39.5) CC/CC 1,130/1,220

B (5.0) AT/AG B (43.5) TC/GC

C (0.5) AA/AA C (17.0) TT/GG

II A (84.5) CC/TT/GG 956/976/980 A (100)

B (5.0) CC/AT/AG

C (10.0) CT/TT/GG

D (0.5) CC/AA/AA

VI A (69.5) _ _/GGc 2655delT/2,678 A (70.0) GG/TT/_ _ 2,922/3,035/

3049ins3050

B (15.0) _T/GT B (24.0) GG/TC/_T

C (14.5) _T/GG C (2.0) GA/TC/_T

D (0.5) TT/GG D (3.5) GG/CC/_T

E (0.5) TT/GT E (0.5) AA/CC/TT

VII A (69.5) _ _/GG 2655delT/2,678 A (70.0) _ _ 3049ins3050

B (15.0) _T/GT B (29.5) _T

C (14.5) _T/GG C (0.5) TT

D (0.5) TT/GG

E (0.5) TT/GT

a Nt: nucleotide positions corresponding to the genotype combination of each DNA fragment, related to the Acc. No. DQ450146b Nt: nucleotide positions corresponding to the genotype combination of each DNA fragment, related to the Acc. No. DQ461643c _: Single nucleotide deletion

SSCP patterns were identified based on the electrophoretic mobility of single-stranded DNA in polyacrylamide gels and were assigned a capital

letter, while genotypes were assigned following analysis of the DNA sequences

Fig. 3 SSCP patterns of GH2-Z

gene DNA fragment I (50UTR),

II (exon-1), VI (exon-5) and VII

(30UTR), A single letter

designation was adopted to

indicate the different patterns.

Each pattern differs from the

others depending on the relative

position of the single strand

DNA bands on polyacrylamide

gel. The genotype combination

corresponding to each pattern is

indicated in Table 2

Mol Biol Rep (2013) 40:5285–5294 5289

123

significantly associated (P \ 0.05) with the highest protein

and casein percentage (Table 6).

Discussion

The separate analysis of the two oGH gene copies,

accomplished in this research, revealed marked differ-

ences: fragment II was monomorphic at the GH2-Z allele,

but showed 4 banding patterns at GH1/GH2-N; fragment

VII had 3 SSCP profiles at GH2-Z vesrus 5 at GH1/GH2-N.

Both GH1/GH2-N and GH2-Z genes showed a higher

number of SSCP patterns at fragments VI and VII

(corresponding to exon-5 and 30UTR), rather than frag-

ments I and II (promoter region and exon-1), similar to that

revealed by Marques et al. [6] in Serra da Estrela sheep.

This confirmed earlier work that indicated the 50UTR and

exon-1 as more conserved than the other regions, at the

GH1/GH2-N allele [7].

Mutations in the promoter region

Prediction of the potential involvement of the nt changes at

the GH1/GH2-N promoter region in transcription factor

binding sites (TFBS) revealed that the two polymorphisms

c.-19T [ A and c.-15G [ A produced a variation

Table 3 Nucleotide changes at

the GH1/GH2-N and GH2-Z

genes in Sarda sheep

a Nucleotide positionb Nucleotide positions are

relative to GH1 genomic DNA

sequence DQ450146c Nucleotide positions are

relative to GH2-Z genomic

DNA sequence DQ461643

Polymorphism Nta Location Deduced

AA change

Frequency

GH1/GH2-Nb c.-39 C [ T 956 50UTR – 0.048

c.-19 T [ A 976 50UTR – 0.030

c.-15 G [ A 980 50UTR – 0.030

c.*30 del T 2655delT 30UTR – 0.843

c.*53 G [ T 2,678 30UTR – 0.076

GH2-Zc c.-259 T [ C 1,130 50UTR – 0.611

c.-169 G [ C 1,220 50UTR – 0.611

c.556 G [ A 2,922 Exon-5 p.Gly186Ser 0.015

c.*15 T [ C 3,035 30UTR – 0.168

c.*29_30insT 3049ins3050 30UTR – 0.151

Table 4 Haplotype frequencies of the GH1/GH2-N and GH2-Z genes in Sarda sheep

Haplotypea c.-39 C [ T c.-19 T [ A c.-15 G [ A c.*30 del T c.*53 G [ T Frequency SEc

GH1/GH2-N

1 C T G _b G 0.830 0.003

2 C T G T G 0.063 0.002

3 T T G T T 0.037 0.001

4 C T G T T 0.029 0.002

5 C A A T G 0.017 0.001

6 C A A T T 0.010 0.001

7 T T G _ G 0.010 0.002

GH2-Z

c.-259 T [ C c.-169 G [ C c.556 G [ A c.*15 T [ C c.*29_30insT

1 C C G T b_ 0.600 0.003

2 T G G T _ 0.232 0.003

3 T G G C T 0.125 0.003

4 T G G C _ 0.018 0.000

5 T G A C T 0.015 0.001

6 C C G C T 0.010 0.003

a Only haplotypes with frequency [0.01 are shown and haplotypes with frequencies [0.05 are in boldb _ = Single nucleotide deletionc SE standard deviations (square root of the variance of the posterior distribution) for the frequencies

5290 Mol Biol Rep (2013) 40:5285–5294

123

Table 5 Analysis of variance showing the effects of oGH genotype and lactation stage (DIM) on milk traits

Fragment II Fragment VI

Milk yield, g/day Fat, % Protein, % Casein, % Milk yield, g/day Fat, % Protein, % Casein, %

GH1/GH2-N

DIM 40.07*** 108.87*** 8.38*** 23.88*** 137.07*** 238.28*** 22.53*** 57.36***

Genotype 4.17* 1.49ns 1.99ns 2.05ns 3.37* 1.77ns 4.54* 3.71*

Genotype 9 DIM 1.69ns 1.30ns 1.10ns 1.26ns 6.52*** 1.79ns 1.64ns 1.66ns

RMSEa 184.85 0.72 0.36 0.30 131.52 0.72 0.35 0.30

GH2-Z

DIM 152.28*** 315.65*** 23.27*** 59.29*** 24.47*** 53.59*** 2.68* 9.68***

Genotype 3.57* 1.87ns 0.39ns 0.53ns 1.08ns 2.02ns 3.30* 3.22*

Genotype 9 DIM 1.67ns 2.28* 0.26ns 0.13ns 1.04ns 0.62ns 0.59ns 0.47ns

RMSEa 134.82 0.71 0.36 0.30 135.25 0.72 0.36 0.30

a RMSE root mean square error

*, **, and *** indicate significant F-values at P \ 0.05, 0.01 and 0.001, respectively

ns not significant

Table 6 LS means of milk traits, according to genotypes at GH1/GH2-N and GH2-Z

GH1/GH2-N GH2-Z

Fragment II Fragment VI Fragment I Fragment VI

SSCP pattern A B C A B C A B C A B C D

N1 169 10 20 139 30 29 79 87 34 140 48 4 7

Milk yield, g/day 586.7ab 581.1ab 454.4b 561.3b 556.7ab 657.7a 617.0a 544.1b 546.8ab 587.7 547.7 564.0 478.2

Fat % 6.40 6.04 6.42 6.44 6.24 6.29 6.37 6.37 6.23 6.41 6.03 5.73 6.74

Protein% 5.89 5.72 6.06 5.94a 5.93ab 5.67b 5.89 5.92 5.84 5.86b 6.01a 5.35c 6.04a

Casein % 4.63 4.50 4.79 4.68a 4.66ab 4.47b 4.64 4.67 4.58 4.61b 4.74a 4.16c 4.76a

a N number of animalsb Different letters in the same row indicate values significantly different; a, b, c = P \ 0.05

Fig. 4 Trends of milk yield, fat,

protein and casein percentages

in Sarda sheep milk during

lactation. Different letters (A, B,

C) differ significantly for

P \ 0.01

Mol Biol Rep (2013) 40:5285–5294 5291

123

depending on the haplotype: the TG combination intro-

duced a PEA3 (Polyoma Enhancer Activator 3) binding

site (nt -14/-23), which was lost when the AA haplotype

occurred, making the nt sequence -17/-26 become a

potential site for the Sp1 (Specificity Protein 1) transcrip-

tion factor. The PEA3 subfamily proteins play key regu-

latory roles in mammary embryogenesis and mammary

gland development in mouse and human [17], while Sp1 is

a transcription factor involved in different aspects of cel-

lular functions in eukaryotic cells [18].

At the GH2-Z promoter region, the occurrence of a C

from the c.-259T [ C base change resulted in a binding

site for C/EBP-alpha (CCAAT/Enhancer Binding Protein)

transcription factor (nt -256/-265), if T substituted for C,

this potential binding site was lost. The C/EBP-alpha

transcription factor is known to have a critical role in

the regulation of adipogenesis [19]. With regard to the

c.-169G [ C polymorphism, the G was involved in a

potential binding site for USF (Upstream Stimulatory

Factor 1) transcription factor (nt -160/-169), which was lost

when C occurred. The USF1 transcription factor is asso-

ciated (in human) to familial combined hyperlipidemia,

and is involved in the regulation of genes for lipid and

cholesterol metabolism [20]. From these data, it can be

observed that there is, within each gene, a variability of the

promoter regions that might have consequences at the

regulatory level on the expression of the related protein.

The two promoter sequences analyzed shared a PEA3

transcription factor binding site at nt -14/-23, and a

region (nt -81/-70) potentially recognized by GATA-1,

Sp1 and AP-2 transcription factors. Upstream of nt -82

sequences of the two promoters and the related predicted

TFBS differ completely. This might explain the differential

expression of the two gene copies pointed out by different

Authors. In fact, Lacroix et al. [10] have detected the

expression of two forms of GH mRNA in the placenta of

sheep, one of which may be related to the pituitary GH, and

has been attributed to the GH1/GH2-N gene. The second

has been attributed to the product of the GH2-Z gene.

Gootwine et al. [9] reported that GH2-Z is not expressed in

the pituitary gland, probably owing to a regulation of gene

expression different from that of GH1/GH2-N.

Mutations in exons

Exon-1 was monomorphic in both GH1/GH2-N and GH2-

Z, as reported by Marques et al. [6] in Serra da Estrela

sheep, suggesting that this region is highly conserved not

only within the GH1 gene copy [7] but also within the

GH2-N and GH2-Z copies.

The nt change detected at the GH2-Z exon-5 has also

been detected in Serra da Estrela sheep [6], and determines

the amino acid substitution p.Gly186Ser of the primary

translation product. The GH residue 186 is located between

helices 3 and 4 of the mature protein; according to some

Authors, the single substitution I186 M in human GH

reduced fivefold its binding to the first receptor, while other

Authors do not report any effect due to substitution of res-

idue 186 [21]. This mutation was analysed with the Panther

[22] and PolyPhen-2 [23] softwares to predict its functional

significance in the gene product. Analysis with Panther

software gave a subSPEC score of -2.87 (higher than -3),

and a P deleterious of 0.467 (lower than 0.5), which means

that probably this coding SNP variant does not affect the GH

protein function, and similar results were obtained with

Polyphen-2. In addition, analysis with NetPhos 2.0 (http://

www.cbs.dtu.dk/services/NetPhos/) server indicated that

the p.Gly186Ser variation does not involve gain or loss of

phosphorylation sites, while analysis with Scansite [24],

when run at a high stringency, hypothesized that Ser186

may introduce a potential PKC_mu phosphorylation motif

at amino acid residues 172–186 of the GH protein.

Mutations in the 30UTR

Several variations were detected at the 30UTR and only one

was common to both GH1/GH2-N and GH2-Z: the inser-

tion/deletion of a Thymine 30 bp downstream the stop

codon (indicated as c.*30delT or c.*29_30insT based on

the reference sequence). An indel at the same position has

also been reported in goat GH gene (GenBank acc. no.

GU355687-9), but with one major difference: the inserted/

deleted nucleotide is a C (not T). Interestingly, a mono-

morphic C occurs at the bovine GH gene at this position.

This site of mutation might bring interesting information

on the phylogeny of Caprines. Moreover, the nt changes

detected in the non-coding regions may affect the regula-

tion of gene expression following the interaction with the

ncRNAs (non-coding RNAs) [25].

The values recorded for milk traits throughout lactation

fall within the range that is typical for the breed (http://

www.assonapa.com/norme_ecc/ovini_llgg/sarda-ovina.htm),

and the fluctuations reflect those of the natural lactation

curve.

At the GH1/GH2-N gene copy, it is interesting to note

the association of pattern A of fragment II and pattern C of

fragment VI with the highest milk yield. These pattern

combinations might be attributed to haplotypes 1 and 2

(Table 4), which were the most frequent, indicating that

these haplotype combinations may have undergone a

positive selection. An influence of fragment VI on the

protein content has been observed at the GH1/GH2-N gene,

which showed the same trend of association with the casein

content, indicating that the influence on the protein content

exerted by some genotypes was due to an influence on the

casein percentage. The correlation of the GH1/GH2-N

5292 Mol Biol Rep (2013) 40:5285–5294

123

copy with the protein percentage was not found by

Marques et al. [6] in Serra da Estrela ewes.

At the GH2-Z gene copy, profile A of fragment I showed

the highest milk yield, it corresponded to the haplotype

combination CC at positions c.-259 and c.-169, respec-

tively. This haplotype combination was involved in a

potential binding site for C/EBPalp transcription factor,

upstream of nt -82, in the region that differs between the

GH1/GH2-N and GH2-Z gene copies. The GH2-Z gene has

been reported to significantly affect milk yield in Serra da

Estrela ewes [6].

Associations of fragment VI with protein and casein

content were evidenced, regardless of the amount of milk

produced. Apparently, the element that differed among

groups within this DNA fragment was the occurrence of an

adenine at the c.556 position (exon-5), corresponding to the

presence of the Ser186 amino acid in the translation

product (pattern C, associated with lower fat, protein and

casein content). So this mutation site might have an

important functional significance, although analysis with

the main softwares did not evidence substantial influences

on the protein function, except for the possible introduction

of a new phosphorylation site.

Associations between polymorphisms within the oGH

gene and milk traits may be due to a direct effect of

pituitary GH gene on the metabolism of lactating sheep, in

relation to the action of the GH1/GH2-N gene copy; or to

the action of GH2-Z gene, expressed in the placenta, which

may exert an indirect effect on milk productions [6]. Also,

it should be taken into account that the effects evidenced

may be due to association of the nucleotide variations

analyzed with causal SNPs, having direct effect on the trait,

yet to be uncovered.

Conclusions

The sheep GH1/GH2-N and GH2-Z gene copies displayed

polymorphic nucleotide changes in the population ana-

lyzed. It was observed, within each gene, a variability of

the promoter regions that might have consequences at the

regulatory level on the expression of the growth hormone,

due to the involvement in potential transcription factor

binding sites.

Both gene copies influenced milk production and com-

position, as some genotypes tended to produce higher milk

yields. These findings expand our understanding about the

sheep GH gene sequence variability and its influence on

milk traits, and may have implications for future strategies

in dairy sheep breeding.

Acknowledgments Research supported by a grant from Regione

Autonoma della Sardegna (L.R.7/2007).

Conflict of interest The authors declare no conflict of interest.

References

1. Etherton TD, Bauman DE (1998) Biology of somatotropin in

growth and lactation of domestic animals. Physiol Rev

78:745–761

2. Lagziel A, Lipkin E, Soller M (1996) Association between SSCP

haplotypes at the bovine growth hormone gene and milk protein

percentage. Genetics 142:945–951 (Corrigendum 1997 Genetics

146 442)

3. Dybus A, Grzesiak W, Szatkowska I, Błaszczyk P (2004) Asso-

ciation between the growth hormone combined genotypes and

dairy traits in Polish Black-and-White cows. Anim Sci Pap Rep

22:185–194

4. Malveiro E, Pereira M, Marques PX, Santos IC, Belo C, Rena-

ville R, Cravador A (2001) Polymorphisms at the five exons of

the growth hormone gene in the algarvia goat: possible associa-

tion with milk traits. Small Rumin Res 41:163–170

5. Marques PX, Pereira M, Marques MR, Santos IC, Belo CC,

Renaville R, Cravador A (2003) Association of milk traits with

SSCP polymorphisms at the growth hormone gene in the Serrana

goat. Small Rumin Res 50:177–185

6. Marques MR, Santos IC, Carolino N, Belo CC, Renaville R,

Cravador A (2006) Effects of genetic polymorphisms at the

growth hormone gene on milk yield in Serra da Estrela sheep.

J Dairy Res 73:394–405

7. Ofir R, Gootwine E (1997) Ovine growth hormone gene dupli-

cation-structural and evolutionary implications. Mammal Gen-

ome 8:770–772

8. Valinsky A, Shani M, Gootwine E (1990) Restriction fragment

length polymorphism in sheep at the growth hormone locus is the

result of variation in gene number. Anim Biotechnol 1:135–144

9. Gootwine E, Ofir YossefiS (1996) Characterization of PvuII

polymorphisms between the ovine growth hormone GH2-N and

GH2-Z gene copies. Anim Biotechnol 7:1–135

10. Lacroix MC, Devinoy E, Servely JL, Puissant C, Kann G (1996)

Expression of the growth hormone gene in ovine placenta:

detection and cellular localization of the protein. Endocrinology

137:4886–4892

11. IZS, Istituto Zooprofilattico Sperimentale dell’Abruzzo e del

Molise ‘‘G. Caporale’’ (2008) Italian National Data Bank for the

animal husbandry register, founded by the Ministry of Health.

http://www.izs.it/IZS/Engine/RAServePG.php/P/256910010304/

L/1. Last accessed 01 Mar 2011

12. Vacca GM, Carcangiu V, Dettori ML, Pazzola M, Mura MC,

Luridiana S, Tilloca G (2008) Productive performance and meat

quality of Mouflon 9 Sarda and Sarda 9 Sarda suckling lambs.

Meat Sci 80:326–334

13. Hall TA (1999) BioEdit: a user-friendly biological sequence

alignment editor and analysis program for Windows 95/98/NT.

Nucl Acids Symp Ser 41:95–98

14. Yeh F, Yang RC, Boyle T (2000) Popgene V.1.32. Microsoft

windows based freeware for population genetic analysis. http://

www.ualberta.ca/*fyeh/popgene.pdf

15. Stephens M, Smith NJ, Donnelly P (2001) A new statistical

method for haplotype reconstruction from population data. Am J

Hum Genet 68:978–989

16. Thorpe DP, Holland B (2000) Some multiple comparison pro-

cedures for variances from non-normal populations. Comput Stat

Data Anal 35:171–199

17. Kurpios NA, Sabolic NA, Shepherd TG, Fidalgo GM, Hassell JA

(2003) Function of PEA3 Ets transcription factors in mammary

Mol Biol Rep (2013) 40:5285–5294 5293

123

gland development and oncogenesis. J Mammary Gland Biol

Neoplasia 8:177–190

18. Li B, Wang X, Zhou F, Saunders NA, Frazer ICH, Zhao K-N

(2008) Up-regulated expression of Sp1 protein coincident with a

viral protein in human and mouse differentiating keratinocytes

may act as a cell differentiation marker. Differentiation

76:1068–1080

19. Payne VA, Au W-S, Lowe CE, Rahman SM, Friedman JE,

O’rahilly S, Rochford JJ (2009) C/EBP transcription factors

regulate SREBP1c gene expression during adipogenesis. Bio-

chem J 425:215–223

20. Rada-Iglesias A, Ameur A, Kapranov P, Enroth S, Komorowski

J, Gingeras TR, Wadelius C (2008) Whole-genome maps of

USF1 and USF2 binding and histone H3 acetylation reveal new

aspects of promoter structure and candidate genes for common

human disorders. Genome Res 18:380–392

21. Juarez-Aguilar E, Castro-Munozledo F, Guerra-Rodrıguez NE,

Resendez-Perez D, Martınez-Rodrıguez HG, Barrera-Saldana

HA, Kuri-Harcuch W (1999) Functional domains of human

growth hormone necessary for the adipogenic activity of hGH/

hPL chimeric molecules. J Cell Sci 112:3127–3135

22. Brunham LR, Singaraja RR, Pape TD, Kejariwal A, Thomas PD,

Hayden MR (2005) Accurate prediction of the functional sig-

nificance of single nucleotide polymorphisms and mutations in

the ABCA1 gene. PLoS Genet 1:e83. doi:10.1371/journal.pgen.

0010083

23. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova

A, Bork P, Kondrashov AS, Sunyaev SR (2010) A method and

server for predicting damaging missense mutations. Nat Methods

7:248–249

24. Obenauer JC, Cantley LC, Yaffe MB (2003) Scansite 2.0: pro-

teome-wide prediction of cell signaling interactions using short

sequence motifs. Nucl Acids Res 31:3635–3641

25. Shimoni Y, Friedlander G, Hetzroni G, Niv G, Altuvia S, Biham

O, Margalit H (2007) Regulation of gene expression by small

non-coding RNAs: a quantitative view. Mol Syst Biol 3:138–146

5294 Mol Biol Rep (2013) 40:5285–5294

123