Skip Navigation


PEDS Advance Access originally published online on January 18, 2008
Protein Engineering Design and Selection 2008 21(3):161-164; doi:10.1093/protein/gzm078
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Discuss on peds methodsblog
Right arrow All Versions of this Article:
21/3/161    most recent
gzm078v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Tobelmann, M. D.
Right arrow Articles by Murphy, R. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tobelmann, M. D.
Right arrow Articles by Murphy, R. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oxfordjournals.org

A strategy for generating polyglutamine ‘length libraries’ in model host proteins

Matthew D. Tobelmann1, Robert L. Kerby2 and Regina M. Murphy1,3

1Department of Chemical and Biological Engineering, University of Wisconsin, Madison, WI 53706, USA 2Department of Bacteriology, University of Wisconsin, Madison, WI 53706, USA

3 To whom correspondence should be addressed: E-mail: regina{at}engr.wisc.edu


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results and discussion
 Funding
 Acknowledgements
 References
 
Huntington’s disease is one of nine known neurodegenerative diseases in which a disease-specific protein contains an unusually long polyglutamine (polyQ) stretch. The proteins associated with each disease are unrelated in sequence, size, structure, function or location of the mutation. In all cases, there is an apparent critical number of glutamines below which individuals do not develop disease. Expansion of the polyQ domain is closely associated with misfolding and aggregation of the protein. It is not yet well understood how the length of the polyQ tract, and its location within a given protein, is related to misfolding and to disease. In this work we developed a strategy for generating length libraries of polyQ-containing proteins, with the polyQ inserted at an arbitrary location. This strategy facilitates systematic, detailed study of the relationship among polyQ length, context and misfolding.

Keywords: aggregation/misfolding/polyglutamine/recombination


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results and discussion
 Funding
 Acknowledgements
 References
 
Protein aggregation is implicated in the pathogenesis of a number of neurological diseases, including β-amyloid peptide in Alzheimer’s disease (Lichtenthaler et al., 2002Go), {alpha}-synuclein in Parkinson’s disease (Gandhi and Wood, 1999Go), prion protein (Prusiner et al., 1998Go) and proteins with expanded polyglutamine (polyQ) domains such as in Huntington’s disease (Wanker, 2000Go). In all nine known polyQ diseases, neuronal inclusions containing aggregates of the polyQ-containing protein or proteolytic fragments thereof are observed (Wanker, 2000Go). Although the diseases are related by phenotype, the disease-specific proteins are unrelated in terms of size, sequence, structure, function and location of the polyQ tract (Wanker, 2000Go). The proteins share the unique phenomenon of an apparent critical polyQ length below which individuals do not develop the disease, i.e. the critical length in Huntington’s disease is 35–40 glutamines but ranges between ~20 in spinocerebellar ataxia 6 and ~60 in Machado–Joseph disease (Gusella and MacDonald, 2000Go). Importantly, the age of onset decreases dramatically as the polyQ tract lengthens. The diversity of disease-specific proteins coupled with the common symptoms and the presence of expanded polyQ domains indicate that the length and location of polyQ plays a defining role in the pathology.

To examine the role of polyQ length in folding and aggegation, investigators have used synthetic peptides (Bhattacharyya et al., 2005Go; Lee et al., 2007Go), and disease-specific proteins including ataxin-1 (Rich and Varadaraj, 2007Go), ataxin-3 (Bevivino and Loll, 2001Go; Masino et al., 2003Go; Shehi et al., 2003Go; Chow et al., 2004Go) and the N-terminal fragment of huntingtin (Scherzinger et al., 1999Go; Wacker et al., 2004Go; Poirier et al., 2005Go). Many of the disease-specific proteins are difficult to express and/or are not structurally characterized. Given that the disease proteins are unrelated except for their common polyQ tract, it has been hypothesized that any protein might serve as an appropriate model host for polyQ. This line of reasoning has motivated studies in which polyQ was inserted into glutathione S-transferase (Masino et al., 2002Go), chymotrypsin inhibitor-2 (Chen et al., 1999Go) and apomyoglobin (Tanaka et al., 2001Go). Interestingly, transgenic mice carrying a non-disease-related protein (hypoxanthine phosphoribosyltransferase), mutated to incorporate a polyQ domain, developed symptoms characteristic of classic polyQ neurodegenerative disease (Ordway et al., 1997Go).

Biophysical studies on the effect of polyQ on the folding and aggregation of a recombinant host protein necessitate the insertion of a new polyCAG DNA sequence, or expansion or contraction of an existing one. In most previous studies, researchers have synthesized only two to four mutants in which the polyQ length was chosen to be above and below the apparent critical length, based on a hypothesis that there is an abrupt and dramatic transition in physical and biological properties as the length changes from non-pathological to pathological. However, more recent evidence suggests that there are not striking qualitative differences between short and long polyQ domains; rather, the conformation and aggregation propensity may change gradually as the polyQ tract lengthens (Klein et al., 2007Go). Furthermore, it has been hypothesized that any length of polyQ could cause disease given a sufficiently long lifespan (Gusella and MacDonald, 2006Go).

We are interested in systematically examining the effect of polyQ length and location on protein folding, stability and aggregation. Practical concerns drove our choice of myoglobin as our host protein. Specifically, a high resolution crystal structure is available, myoglobin is easily produced and (re)-folded and its predominantly helical structure makes following the transition to beta-sheet aggregates easy to detect spectroscopically. Furthermore, Tanaka et al. (2001)Go have successfully used myoglobin as a model for polyQ disease. We required a facile method for inserting varying lengths of polyQ at arbitrary sites within myoglobin, a protein that contains no pre-existing polyQ region. We sought to develop a strategy for generating small libraries of myoglobin mutants containing polyQ domains of incremented, biologically relevant lengths inserted at any chosen site within our host protein.

Several methods for inserting expanded CAG domains into genes and generating length diversity have been reported. Ordway and Detloff (1996)Go employed a polymerase chain reaction (PCR)-based method in which (CAG)20 and (CTG)20 oligonucleotides were subjected to PCR, yielding CAG repeats from 100 bp (~30 glutamines) to 15 kb. This method yielded polyCAG tracts that were much too long for our purpose, and also necessitated the blunt-end ligation of the PCR product into the gene of interest. Takahashi et al. (1999)Go devised a PCR-based method to expand or contract the existing polyCAG regions within a gene. However, this method yielded too little diversity for our application and required a pre-existing polyCAG region. Other methods for expansion of pre-existing polyCAG regions allowed for joining [doubling] an existing polyCAG region (Peters and Ross, 1999Go) or the introduction of stable polyQ-encoding DNA of very specific lengths to pre-existing polyCAG regions (Michalik et al., 2001Go). Lastly, methods exist for synthesizing very long polyCAG tracts (>100 CAG triplets) (Kim et al., 2005Go) flanked with restriction enzyme sites for use in ligation into a gene (Sasagawa and Ishiura, 2006Go). These published methods were not readily adaptable to our application, either because they required an endogeneous polyQ domain, which myoglobin does not possess, or they do not generate the CAG lengths and diversity of interest.

In this paper, we present a novel and simple strategy for site-directed introduction of polyQ-encoding DNA sequences and generation of length libraries. The method generates polyQ libraries with biologically relevant lengths and small increments [2–8 glutamines] between the different lengths. This method will allow us to systematically probe subtle differences among mutants of different lengths.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results and discussion
 Funding
 Acknowledgements
 References
 
Plasmid construction

The sperm whale myoglobin gene in a pETBlue1 expression vector (Novagen, Madison, WI, USA) was a gift from Silvia Cavagnero (University of Wisconsin-Madison, Department of Chemistry). A 6x-histidine tag was added to the C-terminus of the myoglobin gene. Briefly, the gene was PCR-amplified using the primers 5'-GAGAGATCTTGATTGGCTAGC-3' and 5'-TGCGTCGGTACCTCATTAATGGTGATGGTGGTGATGACCCTGGTAACCCAGTTCTTT-3'. The PCR product was then digested with KpnI and NheI (New England Biolabs, Ipswich, MA, USA). A pETBlue1 vector was also digested with KpnI and NheI. Digested DNA fragments (gene and vector) were isolated on a 1% agarose gel stained with 0.5 µg/ml of ethidium bromide and purified using a QIAquick Gel Extraction Kit (Qiagen, Valencia, CA, USA). The histidine-tagged myoglobin gene was ligated into the pETBlue1 vector using T4 DNA ligase (Invitrogen, Carlsbad, CA, USA) and overnight incubation at 16°C. The resulting plasmid was named pMb.

The site of polyQ codon insertion was chosen to be within the flexible CD loop of apomyoglobin, between D45 and K48, leading to the deletion of R46 and F47 (Tanaka et al., 2001Go). A silent mutation was introduced to the R46 codon changing it from CGT to CGA resulting in a novel ClaI restriction enzyme site. The mutagenic primers 5'-CTCTGGAAAAATTCGATCGATTCAAACATCTGAAAACTG-3' and 5'-CAGTTTTCAGATGTTTGAATCGATCGAATTTTTCCAGAG-3' were used in concert with a QuikChange Site-Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA) to perform the mutation. Mutagenesis was confirmed by sequencing.

Site-directed introduction of polyQ codons

The pMb plasmid was digested overnight at 37°C with ClaI (New England Biolabs, Ipswich, MA, USA), then isolated on a 1% agarose gel, stained with 0.5 µg/ml ethidium bromide and purified using a QIAquick Gel Extraction Kit. Two polyQ mutagenic primers were designed to PCR-amplify the entire linear pMb plasmid excluding R46 and F47. To the non-priming (5') end of each primer, a 5'-(CAACAG)8 or 5'-(CTGTTG)8, sequence was added yielding the following polyQ mutagenic primers: 5'-(CTGTTG)8ATCGAATTTTTCCAGAGTTTCC-3' and 5'-(CAACAG)8AAACATCTGAAAACTGAAGC-3'. The sextet repeat CAACAG was chosen to promote polyQ repeat codon stability during protein production (Laccone et al., 1999Go). PCR was conducted using Pfx DNA polymerase (Invitrogen) and the manufacturer’s protocol with the following exceptions: the extention time was changed to 4.5 min and the reaction included ~30 ng linear pMb template and 2.5 U of polymerase. PCR was performed for 35 cycles on a PTC-200 Thermocycler (MJ Research, Waltham, MA, USA).

Ultracompetent MC1061/P3 bacteria (Invitrogen) were used for transformation with 1 µl PCR reaction products. This strain contains the recA gene, unlike most cloning Escherichia coli strains, thus allowing the bacteria to recombine the polyCAA/CAG ends of the PCR products at varying lengths. A 200 µl volume of transformation reaction product was spread on LB agar plates containing 50 µg/ml ampicillin and incubated overnight at 37°C. The following day, single colonies were inoculated in 3 ml LB media containing 50 µg/ml ampicillin and grown in a shaker-incubator at 37°C overnight. The plasmids were purified using a QIAprep Spin Miniprep Kit and subjected to a double restriction enzyme digestion using NheI and KpnI. The DNA fragments from the plasmid digests were separated on a 1% agarose gel stained with 0.5 µg/ml ethidium bromide. Plasmids were sequenced to confirm insertion of polyQ codons.

Protein production and purification

Several of the identified polyQ codon-containing mutated plasmids (named pMb-QXX(CD)) were transformed into TUNER(DE3)pLacI cells (Novagen, Madison, WI, USA). Briefly, single colonies were grown overnight in 100 ml LB media containing 50 µg/ml ampicillin, 34 µg/ml chloroamphenicol and 1% glucose. The overnight culture was used to inoculate 1000 ml of identical medium to an initial optical density (600 nm) of 0.2. The culture was grown into the logarithmic phase (OD600 nm = 0.4–0.7) at which time protein production was induced by adding isopropyl β-D-1-thiogalactopyranoside (Sigma, St. Louis, MO, USA) to 1 mM. The culture was grown to a final OD600 nm of ~2.0. Protein was purified under denaturing conditions using nickel affinity chromatography. Briefly, the cells were centrifuged, the cell pellet was resuspended in 50 ml 100 mM NaH2PO4, 10 mM Tris·Cl, 8 M Urea, pH 8.0, and the suspension was stirred on ice for 30 min to lyse the cells. The lysate was centrifuged at 20 000g for 20 min. The supernatant was collected and loaded onto a gravity-fed affinity column packed with Ni-NTA agarose (Qiagen). Non-specific-binding proteins were washed from the column using the lysis buffer at pH 6.3. The proteins were refolded on the affinity column by rapidly washing the column with PBSA. Pure protein was eluted in PBSA+250 mM imidazole and analyzed by SDS–PAGE on a 4–20% polyacrylamide gel (Pierce Biotechnolgy, Rockford, IL, USA). The samples were boiled for 3 min before loading. Electrophoresis was conducted at 125 V in an XCell SureLock Mini-Cell (Invitrogen) using a running buffer of 100 mM Tris·Cl, 100 mM HEPES and 3 mM SDS at pH 8. Purified horse myoglobin (Sigma) was also run as an additional molecular weight control.


   Results and discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
Funding
 Acknowledgements
 References
 
A schematic depicting our strategy is given in Fig. 1. Briefly, (i) a unique restriction enzyme site was engineered at the desired polyQ insertion site. (ii) The plasmid DNA was digested at the new restriction enzyme site to linearize the plasmid. (iii) After purification, PCR was conducted using mutagenic primers. These primers encode sequences complementary to the ends of the linear plasmid with 5' extensions encoding eight pairs of glutamine sense (CAACAG)8 or anti-sense (CTGTTG)8 codons. (iv) The recA+ E. coli strain MC1061/P3 was transformed directly with the PCR product. Forty-three of the resultant positive transformants were inoculated into LB media with ampicillin and grown overnight. Of these colonies, seven did not grow in liquid media. The plasmid DNA was purified from the remaining 37 colonies and digested. Fragment analysis of the digests showed that 11 plasmids contained either a gene insert of the same size as the wild-type myoglobin (failed polyCAG/CAA introduction) or no plasmid DNA. For the remaining 26 plasmids, the myoglobin gene portion of the plasmid was sequenced and the sequences were analyzed to determine the number of inserted glutamine codons. Table I summarizes these data. Excellent length diversity in the window of interest was obtained. The two most prevalent lengths found were Q16 and Q34. Q16 likely arises due to perfect recombination of the complementary ends of the PCR product. PolyQ inserts of 32 codons likely represent blunt ligation of the PCR product termini, whereas lengths between Q32 and Q16 indicate partial recombination events within the complementary ends. About 20% of the mutants contained tracts longer than 100 codons. We hypothesize that mutants longer than Q32 arise due to either of two mechanisms. First, there may be lengthening of the polyQ codon ends during PCR leading to ends longer than Q16 (i.e. Q20, Q28, etc.) available for recombination after transformation into bacteria. Secondly, mutants initially within the Q16–Q32 range may mutate during growth in recA+ bacteria and gain [or lose] codons, as is commonly known to occur (Ohshima et al., 1996Go; Jakupciak and Wells, 1999Go).


Figure 1
View larger version (11K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. Schematic of the method for site-directed introduction of length libraries of polyCAG/CAA. The steps include (1) engineering of a unique restriction enzyme recognition sequence at the desired polyCAG/CAA insertion site, (2) linearization of the plasmid DNA at the desired insertion site, (3) mutagenic whole plasmid PCR on the linear plasmid using PCR primers with polyCAG/CAA or polyCTG/TTG codons attached and (4) transformation if the PCR product into recA+ E. coli leading to plasmid DNA recombined at various polyCAG/CAA lengths.

 

View this table:
[in this window]
[in a new window]

  		Table I. Length diversity in mutated myoglobin

 
Several plasmids were selected corresponding to myoglobins with polyQ lengths above and below the pathological ‘critical length’ of 35–40 glutamines. The selected proteins were expressed in E. coli and purified by nickel-affinity chromatography. The gel shown in Fig. 2 shows the purity of the protein products and also confirms the relative sizes. The polyQ variant proteins migrate anomalously, with apparent molecular weights that are somewhat higher than the true molecular weight.


Figure 2
View larger version (89K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2. Purified recombinant sperm whale myoglobin and selected polyQ-containing proteins separated on a 4–20% polyacrylamide gel and stained with commassie blue. Purchased horse heart myoglobin was also run as an additional molecular weight control. The lane contents and expected molecular weights are: (1) horse heart myoglobin, 17.1 kDa; (2) sperm whale myoglobin + 6x-His, 18.2 kDa; (3) Q16, 19.9 kDa; (4) Q20, 20.4 kDa; (5) Q24, 20.9 kDa; (6) Q28, 21.4 kDa; (7) Q34, 22.2 kDa; (8) Q38, 22.7 kDa and (9) Q46, 23.8 kDa.

 
Preliminary results indicate that the wild-type myoglobin and the Q16, Q20 and Q24 mutants can be successfully refolded. Circular dichroism spectra indicated that the helical content of refolded sperm whale myoglobin was consistent with its known crystal structure and was similar to spectra obtained with purchased horse myoglobin (not shown). Q16, Q20 and Q24 mutants contained decreasing amounts of helical content as the polyQ tract grew larger. All four proteins bound 1-anilino-8-naphthalene sulfonate similarly (data not shown), suggesting that all retained a heme-binding site. Detailed biophysical studies of these and longer polyQ proteins are underway.

To summarize, we have developed a straightforward method for site-directed insertion of expanded polyCAG/CAA into any gene of interest with generation of a diverse length library. Very preliminary evidence suggests that the method can be further simplified by eliminating steps (1) and (2) shown in Fig. 1: one may be able to directly perform PCR on circular plasmids and obtain the same PCR product used for transformation (not shown). Using this strategy, polyQ codons can be inserted into proteins at regions that contain no endogenous glutamine codons, and a diverse library of length mutants can be generated from a small pool of plasmids. The incremental changes in polyQ length will allow us to closely study the subtle changes in the folding, stability and aggregation of the mutants as the polyQ tract lengthens.


   Funding
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Funding
Acknowledgements
 References
 
National Science Foundation [BES-0330537]; National Institutes of Health [5 T32 GM-08349 to M.D.T., GM-53228 to R.L.K.].


   Footnotes
 
Edited by Lynne Regan


   Acknowledgements
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Funding
 Acknowledgements
References
 
We thank Dr. Darrell McCaslin of the University of Wisconsin Biophysics Instumentation Facility for technical assistance with circular dichroism spectroscopy. We also thank Professor Gary Roberts for several helpful discussions.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results and discussion
 Funding
 Acknowledgements
 References
 
Bevivino A.E., Loll P.J. Proc. Natl Acad. Sci. USA (2001) 98:11955–11960.[Abstract/Free Full Text]

Bhattacharyya A.M., Thakur A.K., Wetzel R. Proc. Natl Acad. Sci. USA (2005) 103:15400–15405.

Chen Y.W., Stott K., Perutz M.F. Proc. Natl Acad. Sci. USA (1999) 96:1257–1261.[Abstract/Free Full Text]

Chow M.K.M., Ellisdon A.M., Cabrita L.D., Bottomley S.P. J. Biol. Chem. (2004) 279:47643–47651.[Abstract/Free Full Text]

Gandhi S., Wood N.W. Hum. Mol. Genet. (1999) 14:2749–2755.[CrossRef]

Gusella J.F., MacDonald M.E. Nat. Rev. Neurosci. (2000) 1:109–115.[Web of Science][Medline]

Gusella J.F., MacDonald M.E. Trends Biochem. Sci. (2006) 31:533–540.[CrossRef][Web of Science][Medline]

Jakupciak J.P., Wells R.D. J. Biol. Chem. (1999) 274:23468–23479.[Abstract/Free Full Text]

Kim S.-H., Cai L., Pytlos M.J., Edwards S.F., Sinden R.R. BioTechniques (2005) 38:247–253.[Web of Science][Medline]

Klein F.A.C., Pastore A., Masino L., Zeder-Lutz G., Nierengarten H., Oulad-Abdeighani M., Altschuh D. J. Mol. Biol. (2007) 371:235–244.[CrossRef][Web of Science][Medline]

Laccone F., Maiwald R., Bingemann S. Hum. Mutat. (1999) 13:497–502.[CrossRef][Web of Science][Medline]

Lee C.C., Walters R.H., Murphy R.M. Biochem. 46:12810–12820.

Lichtenthaler S.F., Beher D., Grimm H.S., Wang R., Shearman M.S., Masters C.L. Proc. Natl Acad. Sci. USA (2002) 99:1365–1370.[Abstract/Free Full Text]

Masino L., Kelly G., Leonard K., Trottier Y., Pastore A. FEBS Lett. (2002) 513:267–272.[CrossRef][Web of Science][Medline]

Masino L., Musi V., Menon R.P., Fusi P., Kelly G., Frenkiel T.A., Trottier Y., Pastore A. FEBS Lett. (2003) 549:21–25.[CrossRef][Web of Science][Medline]

Michalik A., Kazantsev A., Van Broeckhoven C. BioTechniques (2001) 31:250–254.[Web of Science][Medline]

Ohshima K., Kang S., Wells R.D. J. Biol. Chem. (1996) 271:1853–1856.[Abstract/Free Full Text]

Ordway J.M., Detloff P.J. BioTechniques (1996) 21:609–612.[Web of Science][Medline]

Ordway J.M., TallaksenGreene S., Gutekunst C.A., Bernstein E.M., Cearley J.A., Wiener H.W., Dure L.S., Linsey R., Hersch S.M., Jope R.S., et al. Cell (1997) 91:753–763.[CrossRef][Web of Science][Medline]

Peters M.F., Ross C.A. Neurosci. Lett. (1999) 275:129–132.[CrossRef][Web of Science][Medline]

Poirier M.A., Jiang H., Ross C.A. Hum. Mol. Genet (2005) 14:765–774.[Abstract/Free Full Text]

Prusiner S.B., Scott M.R., DeArmond S.J., Cohen F.E. Cell (1998) 93:337–348.[CrossRef][Web of Science][Medline]

Rich T., Varadaraj A. PLoS One (2007) 2:e1014. doi:10.1371/journal.pone.0001014.[CrossRef]

Sasagawa N., Ishiura S. Anal. Biochem. (2006) 357:308–310.[CrossRef][Web of Science][Medline]

Scherzinger E., Schweiger K., Lurz R., Lehrach H., Wanker E.E. Phil. Trans. Royal Soc. Lond. B Biol. Sci. (1999) 354:991–994.[CrossRef]

Shehi E., Fusi P., Secundo F., Pozzuolo S., Bairati A., Tortora P. Biochemistry (2003) 42:14626–14632.[CrossRef][Web of Science][Medline]

Takahashi N., Sasagawa N., Suzuki K., Ishiura S. Neurosci. Lett. (1999) 262:45–48.[CrossRef][Web of Science][Medline]

Tanaka M., Morishima I., Akagi T., Hashikawa T., Nukina N. J. Biol. Chem. (2001) 276:45470–45475.[Abstract/Free Full Text]

Wacker J.L., Zareie M.H., Fong H., Sarikaya M., Muchowski P.J. Nat. Struct. Mol. Biol. (2004) 11:1215–1222.[CrossRef][Web of Science][Medline]

Wanker E.E. Biol. Chem. (2000) 381:937–942.[CrossRef][Web of Science][Medline]

Received November 16, 2007; revised November 16, 2007; accepted November 20, 2007.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Discuss on peds methodsblog
Right arrow All Versions of this Article:
21/3/161    most recent
gzm078v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Tobelmann, M. D.
Right arrow Articles by Murphy, R. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tobelmann, M. D.
Right arrow Articles by Murphy, R. M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?