j. Cosmet.

sci., 55, 327-341 (July/August2004)

Evaluationfor collagenproductsfor cosmeticapplication

YONG PENG, VERONICA GLATTAUER,

JEROME A. WERKMEISTER, and JOHN A.M. RAMSHAW
CSIROMolecularScience,
343 RoyalParade,Parkville,
Victoria 3052, Australia.

Accepted
for publication
January 15, 2004.

Synopsis
Collagenis an importantcomponentfor cosmeticformulation,whereit is an effectivenaturalhumectant
with high substantivity.Commercialcollagenpreparations havea wide rangeof properties.In the present
study, varioustechniqueshave beenusedto examinethree distinct commercialcollagensthat illustrate the
rangeof propertiesthat areavailable.The usefulnessof the varioustechniquesfor assessing
collagenquality
and batch-to-batchvariationis discussed. The resultsindicatethat there are severalsimple, cheap,and
effectivemethods,suchas gel electrophoresis,that provideexcellentinformationon collagenquality. The
appropriateselectionof testsallowsinformeddecisions on the choiceof whichcollagenpreparationto use
to providethe desiredfunctionalityand shelflife of a formulation.

INTRODUCTION

Collagen has becomea valuable and well used componentin cosmeticformulation,

whereit providessignificantbenefits.In particular,it is an effectivenaturalhumectant
(1,2) asa result of the extensive,orderedhydrationnetwork that surroundsthe molecule
(3), in combinationwith its high substantivityto the skin surface.A wide variety of
collagenpreparations is now availablethat aresuitablefor cosmeticformulations.These
preparationsvary considerablyin their compositionand properties.Quality testing to
allow informeddecisionson the choiceof which collagenpreparationto useis therefore
importantsoasto ensurethat the preparationprovidesthe desiredfunctionality.Testing
shouldalsobe capableof monitoringpotentialbatch-to-batchvariationsthat couldalso
affectproductperformance. Also, it shouldindicatewhetherthe collagenis in its native,
triple-helicalform or is presentin its denaturedform, gelatine.
Native collagenis characterizedby a triple-helicalstructure,in which three extended
left-handedpolyprolineII-like helical chainsare supercoiledinto a right-handedtriple
helix, linked throughinterchainhydrogenbonds(4,5). The triple-helicalconformation
requiresa distinctiveamino acid sequence,in particularglycineas everythird residue,

Addressall correspondence
to John A.M. Ramshaw.

327
328 JOURNAL OF COSMETIC SCIENCE

to accommodate the closepackingof the threechains.The extendedconformationof the

individualpolyprolineII-like chainsin the triple helix is stabilizedby the high level of
occurrenceof the amino acids proline and 4-hydroxyproline(Hyp). Hyp, which is
post-translationallymodifiedfrom proline,confersa greaterstabilitythan prolinein the
Y position (6) and is a characteristicamino acid marker of collagen.In tissue,the
collagenmoleculesarecrosslinked to form an extendednetworkthat providesthe tissue
with strengthand durability. These crosslinksare specificand use short, non-triple-
helical peptides, the telopeptides,at the ends of the triple-helical domain of each
collagenmolecule (7).
Bovinecollagenhas beena commonsourcefor cosmeticapplications.However,more
recently,collagensfrom otherspeciesof origin havealsobecomecommerciallyavailable.
Collagensfrom alternativespeciesmay provide a benefit in that potential zoonoses,
particularlytransmissible
spongiformencephalopathies, are lesslikely from sourcessuch
as chickenand fish. However,collagensfrom variousspeciesmay differ in their prop-
erties.For example,they may have different thermal stabilities(8), which could affect
formulationor the shelflife of products.
In the presentstudywe haveexaminedthe suitabilityof a rangeof tests,usinga selection
of commerciallyavailablecollagensamples.Thesesamplesillustratea selectionfrom the
rangeof materialsavailableand allow commentaryon the informationthat eachtest
provides.Sometests,thosefoundin wateranalysisstandards (9), for example,arereadily
availablefrommanyanalyticallaboratories,
whereasothertests,whichmay becriticalfor
providingdiscriminationbetweensamples,may requirethe involvementof more spe-
cialized laboratories.

MATERIALS AND METHODS

COLLAGEN SAMPLES

Commercialcollagensamplessuitablefor cosmeticapplicationwere obtainedfrom the

followingsources: Collasol
© wasfrom CrodaChemicalsLtd (Humberside,UK). CLR
Collagen© wasfrom Chemisches LaboratoriumDr. Kurt Richter Gmbh (Berlin, Ger-
many).AteloHelogen © was a gift from the manufacturer (Meddicoll,North Ryde,
Australia). Solublecollagensampleswere also preparedin our laboratoryfrom fresh,
minceddermis(bovine,porcine,and chicken,obtainedfrom localabattoirs)by pepsin
digestionusing 1 mg/ml pepsin in 100 mM aceticacid adjustedto pH 2.5 with HC1
(10). Solublecollagenfrom diced fish skin (Blue Grenadier(Rlacrz/ro,z/s
sp.) and Nile
perch(Tilapia sp.)obtainedfrom a localsupplier)wasextractedusing0.1 mg/ml pepsin
in 100 mM aceticacid adjustedto pH 2.5 with HC1. All collagenswere purified by
differentialNaC1fractionationat pH 2.5 and then at pH 7.4 (11). Separationof type I
and III collagenswas by rapid ammoniumsulfatefractionation(11).

METHODS

Solutionanalysis.Analysesfor ash,arsenic,and heavymetals(aslead)wereperformedby

a certifiedexternalanalyticallaboratory(MGT EnvironmentalConsultingPty. Ltd.,
Oakleigh, Australia)following standardprocedures(9). pH was determinedusing a
 COLLAGEN EVALUATION 329

RadiometerPHM240 instrument(Copenhagen,Denmark). Conductivity was deter-

mined using a RadiometerCDM3 instrument.
Waterregain.Collagensamplesweredried by lyophilizationin preweighedbottleswith
stoppersand then held oversilicagel for more then 72 h. Sampleswere then stoppered
and reweighed.They were then transferredand openedin an environmentof 32%
constantrelativehumidity, maintainedover saturatedCaC12solution.After equilibra-
tion for morethan48 h, the sampleswerestoppered andreweighed.Water regainby the
dried sampleswas determined,and expressed as a percentageof the dry weight.
Spectroscopy.
ForUV/visiblespectroscopy, spectrawerecollectedfrom400 nm to 650 nm
on a ShimadzuUV-265 recordingspectrometer using cellsof 10-mm path length on
collagensolutionsassuppliedor afterdilution to equalconcentrations
of 1 mg/ml with
milliQ water. For IR spectroscopy,
collagensampleswerelyophilizedand then examined
usinga Perkin Elmer 2000 AutoImagespectrometer with the ATR accessary. Alterna-
tively, samplesmay be analyzedas KBr disks.
E/ectrophoresis.
Collagensamples wereanalyzedby sodiumdodecylsulfate-polyacrylamide
gel electrophoresis(SDS-PAGE) (12), using 5% (w/v) running gels. Prior to analysis,
collagensamplesin samplebufferwere neutralized,if necessary, with 2 M Tris and then
heatedat 100øCfor 2 min. Separationof the otl(I) and otl(III) chainswasby reduction
with 2-mercaptoethanolduring interruptedelectrophoresis (13). Collagensampleswere
alsoexaminedby a non-denaturing,lactic acid buffer system,pH 3.1 (14), using 4%
runninggels.Estimationof the isoelectricpoint, pI, wasby examinationof the direction
of electrophoretic
mobility in gel electrophoresis
in 5% runninggelsin a Tris/BoratepH
8.9 system(15).
Aminoacidanalysis. Lyophilizedcollagensampleswere hydrolyzedin vac•oby constant
boiling (5.8 N) of HC1 containing0.1% phenolin a Waters PicoTagsystemat 108øC
for 20 h. After drying in vac•o,hydrolysates
wereexaminedon a Waters HPLC system
with ninhydrin detection.
Difj•rentialscanning
calorimetry.
Collagenmeltingtemperatures
(Tm)weredeterminedby
DSC using a Nettler Toledo DSC 3000 instrument (Nettler, Schwerzenbach, Switzer-
land). Collagensampleswere preparedat about 5 mg/ml, and equilibratedin 50 mN
aceticacid, pH 2.8, by dialysis.Samples,about 100 mg of solution,were examinedat
a temperatureincrease of løC/min. CollagenT m valuesaregivenasthe temperatureat
the mid-point of the thermal transition.
Scanning electronmicroscopy.
Collagen samplesas provided, or after dilution to equal
concentrations of 3 mg/ml with milliQ water,werecoatedon cleanglasscoverslips and
air driedat roomtemperaturein the cleanair flow of a laminarflow hood.Dried samples
were then gold coatedand examinedin a Philips XL-30 microscope.

RESULTS AND DISCUSSION

GENERAL SOLUTION PROPERTIES

Therearea wide varietyof solutionmeasurements that canbe readilymadeon collagen

preparations
(TableI). Theseincludeanalysisfor toxicmetal components suchasarsenic
and lead,whosepresence wouldmakea preparationunsuitablefor cosmeticapplication.
330 JOURNAL OF COSMETIC SCIENCE

Table I
AnalysisResultson Three Samplesof CollagenSuitablefor CosmeticApplications

Test AteloHelogen
© CLR Collagen
© Collasol
©
Protein content (%w/v) 1.05% 0.28% 4.00%
pH 4.7 3.8 4.2
pI >8.9 >8.9 <8.9
Arsenic < 1 ppm < 1 ppm < 1 ppm
Heavy metals <5 ppm <5 ppm <5 ppm
Dry weight 1.15% 5.07 % 5.25 %
Conductivity 0.45 ms 19.0 ms 42.0 ms
Ash content <0.1% 1.2% 1.1%
Hydration regain 21% 3% 7%

In the presentstudy,the levelsof thesecomponentswere below the detectionlimit for

all three samples.
An importantfeatureof a collagenproductis its collagenproteinconcentration. For fully
solubleand essentiallypure collagenpreparationswith concentrations aboveabout 2-3
mg/ml (or solutionsafter clarificationby centrifugation),a Biuret assay(16) against
appropriatecollagenstandardsis simpleand rapid. However,non-collagenous impuri-
ties could interfere with this determination,leading to higher values.Total nitrogen
content can also be readily determined by Kjeldahl analysis.This value allows the
collagenconcentration of the sampleto be estimated,usinga conversion factorbasedon
the knownstructureof collagens(17). This factoris dependenton the collagentypeand
the speciesof origin. Again, this methodcangive erroneous valuesfor collagencontent
if any non-collagenous proteinsare present.Hydroxyprolineis found as an amino acid
that is characteristic
of collagenousproteins,and thereforeprovidesa specificmethodfor
determinationof collagencontent,separatefrom any otherproteinsthat may be present.
Hydroxyprolinecontentand the total amino acid compositioncan be determinedby
amino acid analysis(seebelow). Hydroxyprolinecontentcan also be determinedby a
specificcolorimetricassay(18).
The useof dry weight is not suitableto determinecollagencontent,assomepreparations
containa significantsalt content,ascanbe seenfrom the ashcontentof samples(Table
I). The conductivityof a samplecanconfirmthe presence of salts(Table I). For example,
AteloHelogen
© showsa negligibleashcontent,andhasa verylow conductivity,
whereas
both CLR Collagen
© and Collasol
© both havea significantashcontentand significant
conductivities,althoughthesemeasures arenot necessarilylinked, asthey dependon the
nature of the saltspresent(Table I).
The pH valuesfor the preparations examinedin this studyvaried(TableI). Thesevalues
may be importantin developingformulations.The presence of saltsmay affectthe pH,
astheymaybebufferingthe collagensolution(TableI). Samples of collagenthat contain
oligomersare moresolubleunderlow pH conditions--at higherpH the collagenoligo-
mersbecomelesssolubleand may precipitate(10). On the other hand,at a higherpH,
suchasneutralpH, a minimum saltcontentof around0.15 M NaC1is neededto ensure
collagensolubility(10). Thus,the higherpH of the AteloHelogen
© collagenwith a very
low salt content (Table I) indicatesa highly monomericcollagenpreparation.
WATER REGAIN

The intrinsicwater-bindingpropertyof collagenis criticalto its excellentperformance

 COLLAGEN EVALUATION 331

asa humectantin cosmeticpreparations. An estimateof thiswater-bindingcapacitymay

be obtainedfrom the water regain by dry samplesheld under constanthumidity (32%)
(Table I). It can be seenthat all collagensamplesin the presentstudy showedwater
regainafter drying, but the extent or rate varied betweenthe samplesfor the present
method at 32% relative humidity. When sampleswere held at 87% relative humidity
(saturated
Na2CO3),all absorbed
sufficientwater to form solutionsor wet slurriesof
collagen(data not shown).The variationbetweensamplescouldreflectthe presenceof
salts.The conductivity
andashcontentdata(TableI) showthat AteloHelogen
© hasa
very low salt content,whereasthe other two samplesboth containsalts.The natureof
the salts(which wasnot determined)couldappearto increasethe collagenwater regain
if they are particularlyhygroscopic,but this is not apparentin the presentstudy.
Alternatively,they could reducethe water regain,by slowingdown the rate at which
equilibriumis attained,for example.An alternativeapproachto comparesampleswould
be to examinethem after extensivedialysisto removeany saltspresent.

SPECTROSCOPY

Commercialcollagenpreparations areusuallycolorless solutionsor theymay be white if

there is any insolublematerial present.The principal use, therefore,of UV/visible
spectroscopy is not to detectcoloredimpurities,but ratherto assess the levelof turbidity
that may be presentin a solublecollagenpreparation.For clearsamples,it may alsobe
usedto detectand potentiallyquantirateany UV-absorbingpreservatives that may be
present.Materialmay be assessed as suppliedor after dilution, for example,either in
water so as to minimize changesin pH or in dilute aceticacid so as to ensurethe
solubilityof non-cross-linked components, as in certainpreparationsit is possiblethat
solublecollagenhasformedinto fibrils at neutralpH.
Asanillustration,comparison
of the threetestcollagens
showsthatAteloHelogen
© was
particularlytransparent
assupplied,whereas bothCLR Collagen© andCollasol
© were
turbid (Figure1A). After dilutionin waterto equivalentconcentrations
of 1 mg/ml, the
clarityof Collasol
© wasstill poor,while that for CLR Collagen© wassignificantly
improved,althoughstill muchlessthanAteloHelogen © (FigurelB). The solubilityof
Collasol
© wasimprovedin aceticacid (datanot shown).
IR spectroscopy(Figure2) providesa methodfor showingproteinidentityin the sample.
It mayalsoshowthe presence of organicbuffersor preservatives.
IR spectroscopyalsohas
potentialfor assessing
the extentof collagendenaturationin a collagenpreparation,as
anygelatinepresent
changestherelative
intensity
ofthebandsat 1660cm-• and1633
cm-• (19).However,
thisrequires
highresolution.
In thepresent
study,theresolution
of the IR spectrawasnot sufficient,andalsoseemedto varybetweenpreparations(Figure
2). ThusIR spectroscopy maynot be a convenient methodfor estimatingthe contentof
denatured collagen(gelatine)presentin a sample.The collagen/gelatine
contentof fully
clearsolutionsmayotherwisebe examinedby ORD or CD spectroscopy, wherestandard
valueshave been reported(20,21), but this requiresaccessto specializedequipment.
Sampleclarityis an issue:insolublecollageninterferes
andcanbe removedby centrifu-
gation, but the resultinganalysisis not representative
of the total sample.Lossof
gelatinecomponents afterbriefproteolysis,
wherecollagenis stable,followedby colla-
genprecipitation,alsoprovidesa convenient methodfor estimatinggelatinecontent.
332 JOURNAL OF COSMETIC SCIENCE

A 100%

50%

•3

2
0%
400 nm 500 nm 600 nm

100%

50% 3

0%
400 nm 500 nm 600 nm

Wavelength
Figure 1. UV/visiblespectrafor threecollagens
for cosmetic
use:(1) AteloHelogen
©, (2) Collasol
©, and(3)
CLRCollagen
©at theconcentration
supplied
(A) andafterdilutionto equalconcentrations
of 3 mg/mlwith
water (B).

ELECTROPHORESIS

Variouselectrophoretictechniques are suitablefor collagenanalyses.

Thesetechniques
arecheapand readilyperformed,particularlyif commercially availablepre-castgelsare
used.In the examples,(Figures3-5), the threedifferentcollagensfor cosmeticusehave
beencompared. Sincethe stainthat isusedto detectthe collagenouscomponents in each
of the electrophoretic
methodsis not collagen-specific but stainsall proteins,these
electrophoretic
methodsalsoallow any protein contaminants(serumalbumin, for ex-
ample)to be detectedin collagensamples.
In SDS-PAGE (Figure 3), the molecularweight distribution of the individual chain
components afterdenaturation
is shown.The single-chaincomponents,the o•-chains,
are
the fastestmoving components from intact collagens.Thesechainsshowthat Atelo-
Helogen
© contains
typeIII collagen,
whilethisisabsentfromCLRCollagen
©.Collasol
©
showsan atypicalchainpattern,whichalsopreventsassessment
of anytypeIII collagen
present.None of the collagensshowany significantbandsmoving fasterthan the
o•-chains;
if present,thesebandscouldindicatedegradationproducts,althoughthey
 COLLAGEN EVALUATION 333

%T

Figure 2. Infra-redspectrafor threecollagens

for cosmetic
use:(1) AteloHelogen
©, (2) Collasol
©, and(3)
CLR Collagen ©.

may alsooccurif the sampleis heatedin samplebufferfor >2 min prior to electropho-
resis.Crosslinkedchaincomponents, the dimer [3- and trimer 'y-chains,for example,
movemoreslowlyon the gel. Theseindicate(Figure3) that AteloHelogen
© hasfewer
crosslinked
components than the othercollagens.The bandsthat are presentare those
normallyshownby purifiedtypeI andtype III collagens
(22), and no bandsdueto other
collagentypesareevident.Again,the patternshownby the Collasol
© is complexand
atypical,with a high proportionof crosslinked
material,muchof which is too largeto
enter the gel.
Electrophoresiscanalsobe performedundernon-denaturingconditions,and useshighly
porousgelsto allow the largercollagenmoleculesto migrate(14). This system(Figure
4) showsthe solublecollagenthat is presentin a preparation.Collagendimersalsoenter
the gel, but insolubleor highly polymericmaterial doesnot enter. In the example
(Figure4), it canbeenseenthat all three samplescontainsolublemonomericcollagen
of nativesize.AteloHelogen © showsfewer crosslinked dimer components than CLR
Collagen
©. Collasol
© showsa poor,streakypattern,with lesscollagenbehavingas a
solublecomponent. This suggests that muchof thiscollagenmaybe crosslinkedor have
been modified during preparation.If the isoelectricpoint is low (see below), then
material may not enter the gel as readily at pH 3.1.
Gel permeationchromatography can also be usedto examinethe molecularweight
distributionof solublematerialin a collagenpreparation(23). However,this requires
specializedequipmentand columns.Also, as all insolublematerialmust be removed
prior to analysisso that it doesnot block the flow in the column,the data are not
representativeof the completesample. Electrophoresis providesa simpler, better
method.
334 JOURNAL OF COSMETIC SCIENCE

Origin

Cross-linked Trimers

Cross-linked Dimers-

Monomer (zl [111]

__
Monomer
(zl[I]_ • ,.........
•
Monomer •211]_ :

1 I 2 2 3 3
Figure3. SDS-polyacylamide
gelelectrophoresis
ofthreecollagens
forcosmetic
use:(1) AteloHelogen
©,(2)
Collasol
©, and (3) CLR Collagen
©.

Electrophoresis
canalsoprovidea simplemethodto estimatethe isoelectric
point of the
collagenpreparation.Native collagenshavea high pI, typicallyabovepH 9 (24),
consistentwith the determinedaminoacidsequence
data(17), whereascollagenthat has
beenpreparedfrom alkali-treatedmaterial(as a by-productof liming during leather
making,for example)will havebeendeamidated andwill havea low pI value(24).
The migrationof nativecollagen,towardsanodeor cathode,is an indicationof its net
chargeat a givenpH. For example(Figure 5), collagensthat migratewith a net positive
chargeat pH 9.0 havehighpI valuestypicalof nativecollagens.
Thus,the presentdata
indicatethat bothAteloHelogen
© andCLRCollagen © havethehighpI valuesfoundfor
native,unmodifiedcollagen,whereasCollasol© showsa low pI, suchasmay be found
from lime-treatedhide collagen.

AMINO ACID ANALYSIS

Aminoacidanalysis dataarepresented for the threedifferentcollagensamples(TableII).

This techniqueis readilyavailablethroughanalyticalserviceprovidersif it is not
availablein-house.Thesedata showthat all sampleshavethe high Gly content,around
one third of all amino acids,that is a characteristicof collagens.If a large amount of
proteinimpurityispresent,thisvaluemaybecomelower,but formostsamples thistype
of analysiswill not allowaccuratedeterminationof smallquantitiesof impurity. These
 COLLAGEN EVALUATION 335

Cross-linked Dimers

I I 2 2 3 3
Figure4. Native(lacticacidpH 3.1)polyacrylamide
gelelectrophoresis
ofthreecollagens
forcosmetic
use:
(1) AteloHelogen
©, (2) Collasol
©, and(3) CLR Collagen
©.

dataalsoshowthe high contentin Pro andHyp, aroundonefifth of all aminoacids,that

are essential
elementsof the collagenstructure.The aviancollagen,AteloHelogen ©,
whichis morethermallystable(seebelow),showsan increased Pro and Hyp content.
The amount of Tyr presentshowsthe extent to which the telopeptideshave been
removed,as this aminoacidis generallyonly foundin theseregionsin type I collagen.
However,type III collagendoescontaina low level of Tyr within the helical domain
(17); if analysisshowsthat this collagentype is present,this mustbe takeninto account
(seebelow).Of the three collagensexamined,AteloHelogen ©, which is describedas
monomeric,showsa very low Tyr content,despitethe type III collagenshownby
electrophoresis (Figure 3).

Amino acidanalysisalsoprovidesan approach

to quantitationof the collagenprepara-
tions.Thesedatamaybe obtainedby usingthe Hyp content,andrelatingthis valueto
theknownamountof thisaminoacidin thespecies of thecollagen
(17). Alternatively,
theyieldsof all aminoacidscanbedetermined
andsummed.A keyissuein determining
collagencontent is the water contentthat is present.Certain water moleculesform an
integralpart of the structureand cannotbe readilyremovedfrom nativecollagens by
drying.Other waterpresentis morelooselybound;the amountof this wateris depen-
denton the relativehumidity of the environmentandsopresentssignificantdifficulties
in collagenquantitationif weighingof dry samplesis needed.The useof amino acid
analysisallowsa direct measureof the collagencontentof a collagensolution.
336 JOURNAL OF COSMETIC SCIENCE

1 2 3 1 2 3
pH 8.9, Toward Cathode(-) pH 8.9, Toward Anode(+)
Figure 5. EstimationofpI for threecollagensfor cosmeticusefrom electrophoretic
mobility at pH 8.9: (1)
AteloHelogen©, (2) Collasol
©, and(3) CLR Collagen ©.

DIFFERENTIAL SCANNING CALORIMETRY

Variousmethodshavebeendescribedfor determiningthe thermalstabilityof collagens,

wherethe changein a parameteris followedwith increasingtemperature.Theseinclude
spectralmethods, such as UV spectroscopy (25), ORD (20), and CD (21) enzyme
digestionsusceptibility(26), and calorimetricmethods,suchas differentialscanning
calorimetry(DSC) (27,28).
The resultsobtainedcandependon the methodused.For example,evenfor any single
method,differentresultsmay be obtainedby changingthe rate of heating.Also, the
methodusedfor calculatingand reportingthermalstabilityis important.For example,
mostcorrectlyreportthermal stabilityas the midpoint of the thermal transition,as in
the presentreport. However, others,typically measuringpreparationswith lower sta-
bility, may report the higher temperatureat the end of the transition.Thus, thermal
stabilitymust be comparedusingdefinedstandardconditions.
In the presentstudywe haveexaminedthe thermal stability of a rangeof collagens,
includingthoseavailablefor cosmeticformulation,by a singlemethod,DSC, using
aceticacid as the solventto ensuresamplesolubility.This methodwas chosenbecause
it is generallyreadilyavailableand hasprovenreproducibility.Data arereportedusing
the generallyacceptedmethod of the midpoint of the thermal transition.
All collagensgave distinct single endothermictransitionsby DSC (Figure 6), with
similar transitionenergiesfor comparablequantitiesof collagens.The thermal transi-
 COLLAGEN EVALUATION 337

Table II
Amino Acid AnalysisResultsfor Three Different CollagenPreparations

Amino acid AteloHelogen

© CLR Collagen
© Collasol
©
HOPro 98 73 83
Asp 42 48 44
Thr 17 19 17
Ser 22 31 28
Glu 73 80 80
Pro 125 121 123
Gly 347 325 342
Ala 114 112 113
Cys n.d. n.d. n.d.
Val 19 25 26
Met n.d. n.d. n.d.
Ile 11 14 13
Leu 24 31 29
Tyr 1 7 3
Phe 13 16 14
His 4 6 4
HOLys n.d. n.d. n.d.
Lys 28 31 26
Arg 54 55 50
Trp n.d. n.d. n.d.

Data are expressed

as residuesper 1000. HOPro: 4-hydroxyproline.HOLys: hydroxylysine.
Amino acidssensitiveto oxidation(Met and Cys) were not determined.
n.d.: not determined.

tionsallowedthe determinationof T mvalues(TableIII). Replicatedeterminations

were
reproducibleto 0.3øC or better.
Significantvariationin melting temperaturewas evidentamong the collagensfrom
differentspecies(TableIII). Comparison of typeI collagenT mvaluesshowsthat chicken
collagenwas the most stableat 44.7øC, about 2.5øC more stablethan the mammalian
samples,whichhadT mvaluesin a rangebetween42.0øCand42.2øC.The fishcollagens
were considerablylessstable(Table III), consistentwith the previousobservationthat
collagenstability approximates the upperenvironmentaltemperatureexperiencedby an
organism(8). Thus Macruronus livesin coldwaters(29), giving it a lowerenvironmental
temperaturethan Tilapia, which lives in tropical regionswherefarmedfish may have
environmentaltemperaturesaround 30øC (30). The Tilapia collagen,which has been
suggested assuitablefor cosmetic applications,
gavea T m of 35.8øC(TableIII). There
is somevariationof T mwith the pH of the samples,andsothe pH of cosmeticcollagen
samplesneedsto be consideredif the solutionsare not made up in a standardsolvent.
The melting data showthat for a given species,the type III collagenhad a slightly
greaterT m value, about 1.2øC higher for mammalsand 0.4øC higher for chickens,
consistentwith the stabilizing effectsof a higher hydroxyprolinecontent and the C-
terminaldisulfidebondsthat arepresentin type III collagen(17,31). Althoughpurified
chickentype I and type III collagensshowa smalldifferencein T m (Table III), Atelo-
Helogen ©, which containsa low level,about7-10% of type III collagen(Figure3,
personalcommunication, Meddicoll),gavea singletransitioncomparable to the purified
chickentype I collagen,and a highertransitiondue to the type III componentwasnot
observed(Figure 6).
338 JOURNAL OF COSMETIC SCIENCE

'1 ................... '''1'- -'. ..... ':''t ' ' -•'

Temperature( øC;)
Figure 6. The DSC meltingcurvefor AteloHelogen
©, showinga meltingpoint determinedfrom the
midpoint of the transitionat 44.6øC.

Table III
Melting TemperatureValuesDeterminedby DSC for VariousCollagens

Collagentype To• pH 2.8 (øC)

Bovine I 42.2
Bovine III 43.5
Pig I 42.0
Pig III 43.2
Chicken I 44.7
Chicken III 45.1
Nile perch (Tilapia sp.) 35.8
Blue Grenadier (Macr•ronm sp.) 23.0
AteloHelogen
© 44.6

Thus,for monomericcosmetic collagens, the aviancollagen,AteloHelogen

©, wouldhave
a greaterstabilitythan the bovinecollagens,which may providesignificantbenefitsfor
formulation and shelf-life stability. Tilapia collagenhas been suggestedfor cosmetic
applications,but it has a lower stability than the bovineand particularlythe avian
collagens.

SCANNING ELECTRON MICROSCOPY

SEM providesan approachto evaluatingany fibrousor particulatematerialthat may be

presentin a collagenpreparation.The samplesweredrieddownontoa cleanfiat surface
forevaluation.In the examples
(Figure7), theAteloHelogen© (panel1), whichpresented
 COLLAGEN EVALUATION 339

Figure 7. SEM imagesof thin filmsof threecollagens

for cosmetic
use:(1) AteloHelogen
©, (2) Collasol
©,
and(3) CLR Collagen
©. Scalebarsareshown.

as a clearsolution,showedan essentiallyfeatureless
film, confirmingthe lack of par-
ticulateor fibrouscomponents.On the other hand, the Collasol©, which wasa white
turbid sample,showedextensivefibrousstructures(Figure7, panel2). Of interestwas
theCLRCollagen © sample,which,althoughpresenting
asa fairlyclearsolution,showed
a rangeof blackor greyzonesunderSEM (Figure7, panel3). The natureof thesespots
is not clear.

OTHER TECHNIQUES

A rangeof other techniquesmay also be valuablein assessing collagenfor certain

applications.
Theseincludedeterminationof viscosityandsolubility,overa rangeof pH
valuesto matchformulationrequirements;
density;refractiveindex;and lipid content.
Microbiologicalcontentand immunologicalresponse may alsobe issuesof concern.

CONCLUSIONS

The collagensamplesexaminedin the presentstudy show a wide range in various

340 JOURNAL OF COSMETIC SCIENCE

properties.This rangesuggests that oneof thesecollagenscouldprobablybe selectedto

meet specificmanufacturingneeds.The teststhat havebeendescribedallow the poten-
tial collagensto be evaluated,and differentbatchesto be comparedfor quality assurance.
The teststhat shouldbe applied to sampleswill dependon the needsof the particular
formulationbeingdeveloped. A keypropertyof collagens, their water-bindingcapacity,
may alsobe measured,but this dataneedsto be treatedwith care,asthe saltpresentin
certainpreparationscanlead to falseinterpretations.Of the collagensexamined,Atelo-
Helogen
© wasmonomeric
andparticularly
pure,bothbiochemically
andin theabsence
of salts.

The country of origin may be a considerationat presentdue to risk of any viral or

prion-basedcontamination. Thus, non-bovinecollagensmay alsobe preferredasthere
may be lessrisk of diseasetransmission. In this respect,the chickencollagen,Atelo-
Helogen ©, is intrinsicallymorestablethan marinecollagens. Consumer sensitivityto
avian collagensis potentiallylow, and comparableto mammaliancollagensdespitethe
structuraldifferences(data not shown).It hasalsobeensuggested that collagencould
permeateintact skin and augment the collagenoustissue.This seemsimplausible,
however,as collagenmoleculeswould be too large and too well bound to penetratethe
stratum corneum(32). Should collagenor fragmentspassinto the underlyingtissue,
they would be unableto participatein the complexbiosynthetic
pathwaythat charac-
terizescollagendepositionin tissue(7). Evidencehasbeenpresentedthat confirmsthat
this augmentationdoesnot occurin skin with intact stratum corneum(33). Indeed, the
lack of skin penetrationshouldminimizesensitivityissues.

ACKNOWLEDGMENTS

We thank N. Bartonefor assistancewith aminoacid analysis,G. Heath for assistance

with IR spectroscopy,
S. McCarthy for assistance
with DSC, and Dr J. Ward for assis-
tance with scanningelectron microscopy.We thank our suppliersfor the gifts of
collagenpreparations.

REFERENCES

(1) P. Morganti,S. D. Randazzo,and A. Cardillo,Role of insolubleand solublecollagenas skin mois-

turizer,J. AppLCosmetoL,4, 141-152 (1986).
(2) J. A.M. Ramshaw,Collagen,Cosmet. AerosohToilerr.,1, 22-25 (1986).
(3) J. Bella, B. Brodsky,and H.M. Berman,Hydration structureof a collagenpeptide,Structare, 3,
893-906 (1995).
(4) R. D. B. Fraser,T. P. MacRae,andE. Suzuki,Chainconformation
in the collagenmolecule,
J. Mol.
BioL, 129, 463•481 (1979).
(5) J. Bella,M. Eaton,B. Brodsky,andH. M. Berman,Crystalandmolecularstructureof a collagen-like
peptide
at 1.9• resolution,
Science,
266,75-81 (1994).
(6) R. A. BergandD. J. Prockop, The thermaltransition ofa non-hydroxylated formofcollagen.
Evidence
for a rolefor hydroxyproline
in stabilizingthe triple-helixof collagen,Biochem.
Biophys.
Res.Cornman.,
52, 115-120 (1973).
(7) J. F. Bateman,S. Lamand•,andJ. A.M. Ramshaw,"CollagenSuperfamily,"in Extracellalar
Matrix:
Vol. 2. MolecalarComponents
and Interactions,
W.D. Cornper,Ed. (HarwoodAcademicPublishers,
Amsterdam,1995), pp. 22-67.
(8) B.J. RigbyandM. S. Robinson,Thermaltransitionsin collagen
andthe preferred
temperature range
of animals, Nature, 253, 27-279 (1975).
 COLLAGEN EVALUATION 341

(9) StandardMethods
for theExamination
of Waterand Wastewater
(AmericanPublic Health Association,
American Water Works Associationand Water Environment Federation,Washington, D.C., 1995).
(10) E.J. Miller andR. K. Rhodes,Preparation andcharacterizationof the differenttypesof collagen,Meth.
Enzymol., 82, 33-64 (1982).
(11) R. L. Trelstad,V. M. Catanese, and D. F. Rubin, Collagenfractionation:Separationof nativetypesI,
II andIII by differentialprecipitation,Anal. Biochem.,
71, 114-118 (1976).
(12) U. K. Laemmli,Cleavage of structuralproteinsduringthe assembly of the headof bacteriophage T4,
Nature, 227, 680-685 (1970).
(13) B. Sykes,B. Puddle,M. Francis,and R. Smith,The estimationof two collagens
fromhumandermis
by interruptedgel electrophoresis,
Blochem.Biophys.
Res.Commun., 72, 1472-1480(1976).
(14) J. A.M. Ramshawand J. A. Werkmeister,Electrophoresis and electroblottingof native collagens,
Anal. Blochem.,168, 82-87 (1988).
(15) T. McLellan,Electrophoresis
buffersforpolyacrylamide
gelsat variouspH, Anal.Biochem,
126, 94-99
(1982).
(16) A. G. Gornall,C.J. Bordawill,andM. M. David, Determinationof serumproteinsby meansof the
biuret reaction,J. BioLChem.,177, 751-766 (1949).
(17) K. Kadler, Extracellularmatrix. 1. Fibril-formingcollagens,ProteinProdq/e,5, 519-638 (1994).
(18) G. Huszar,J. Maiocco,and F. Naftolin, Monitoring of collagenand collagenfragmentsin chroma-
tographyof proteinmixtures,Anal Biochem., 105,424-429 (1980).
(19) K.J. Payneand A. Veis, FouriertransformIR spectroscopy of collagenand gelatin solutions:Decon-
volutionof the amideI bandfor conformational studies,Biopo/ymers,
27, 1749-1760 (1988).
(20) M. NagelschmidtandB. Viell, Polarimetricassayfor the determination of the nativecollagencontent
of solublecollagen,J. Biotaed.Mater. Res.,21,201-209 (1987).
(21) T. Hyashi,S. Curran-Patel, andD.J. Prockop,Thermalstabilityof the triple helix of typeI procol-
lagenand collagen:Precautions for minimizingultravioletdamageto proteinsduringcirculardichro-
ism studies,Biochemistry,18, 4182-4187 (1979).
(22) J. A.M. Ramshaw,Distributionof type III collagenin bovineskin of variousages,Connective Tissue
Res.,14, 307-314 (1986).
(23) R. A. Condell, V. P. Hanko, E. A. Latehas,G. Wallace, and K. A. McCullough, Analysisof native
collagenmonomers andoligomersby size-exclusionhigh-performance liquid chromatographyand its
application,Anal. Blochem.,
212, 436•i45 (1993).
(24) J. E. Eastoeand A. A. Leach,"ChemicalComposition of Gelatin,"in TheScienceand Technologyof
Gelatin, A. G. Ward and A. Courts, Eds. (AcademicPress,London, 1977), pp. 73-107.
(25) C. C. Danielsen,Precisionmethodto determinedenaturationtemperatureof collagenusingultraviolet
difference
spectroscopy,
Coll.Relat.Res.,2, 143-150 (1982).
(26) P. Bruckner,and D.J. Prockop,Proteolyticenzymesasprobesfor the triple-helicalconformation
of
procollagen,
Anal. Blochem.,
110, 360-368 (1981).
(27) W. Friessand G. Lee,Basicthermoanalytical
studiesof insolublecollagenmatrices,Biomateriah,17,
2289-2294 (1996).
(28) D. G. Wallace,R. A. Condell,J. W. Donovan,A. Paivinen,W. M. Rhee,and S. B. Wade, Multiple
denaturational transitionsin fibrillar collagen,Biopolymers,
25, 1875-1895 (1986).
(29) J. A.M. Ramshaw,J. A. Werkmeister,and H. A. Bremner,Characterization of type I collagenfrom
the skinof bluegrenadier(Macruronus novaezelandiae),Arch.Blochem.Biophys.,267,497-202 (1988).
(30) H. C. Wang, J. D. Chen,G. C. Li, and F. H. Yew, Characterization of a heat-resistant
strainof Tilapia
ovarycells,J. CellSci.,92, 353-359 (1989).
(31) N. K. Shah, M. Sharma,A. Kirkpatrick, J. A.M. Ramshaw,and B. Brodsky,Gly-Gly-containing
tripletsof low stabilityadjacentto a typeIII collagenepitope,Biochemistry, 36, 5878-5883 (1997).
(32) M. Chvapil,Collagenin cosmetics: Misconception of its modeof action,Cosmet. Toiletr.,97, 35-36
(1982).
(33) S. D. Coapman,J. L. Lichtin, A. Sakr,andJ. R. Schiltz,Studiesof the penetrationof nativecollagen,
collagenalphachains,and collagencyanogenbromidepeptidesthroughhairlessmouseskin in vitro,
J. Soc.Cosmet.
Chem.,39, 275-281 (1988).