1

Analysis of organic components in the test wall of agglutinated foraminifera by Fourier Transform Infrared and Pyrolysis Gas Chromatography!

Mass Spectrometry

K1THRYN ALLEN, STEPHEN ROBERTS and JOHN W.A1URRAY x-C-,,:-.:?l oi Ocean and Earth Science, Southampton Oceanography Centre, Empress Dock, European Way, ~:champton, S014 3ZH, UK.

ABSTRACT Fourier Transform Infrared spectroscopy and Pyrolysis Gas Chromatography /Mass Spectrometry are techniques that reveal specific information on the nature of organic materials in the test walls of agglutinated foraminifera. Organic linings and tests from agglutinated foraminifera Astrammina rara Rhumbler, Jadammina macrescens (Brady), Troclwmmina illJIata Montagu, and additionally the lining from calcareous foraminifera Ammonia beccarii (Linne) were isolated and analysed. Analysis of reference materials, chitin and collagen, enabled comparison with organic components in foraminifera test walls. Absorptions within FT-IR spectra highlight the presence of protein, protein secondary structures, and specific amino acid and carbohydrate components present in foraminiferal tests, linings and also reference materials. Spectral enhancement techniques were used to reveal detail about the secondary structure of the proteins from lining and shell materials. Pyrolysis GC/MS results indicate that protein is the main constituent, with a smaller carbohydrate component in all foraminiferal organic linings and shells. Results from both analytical techniques indicate that foraminiferal linings and cements are glycoproteinaceous materials. Similarity was striking between the FTIR spectra, and pyrolysis GC/MS products from both collagen and foraminiferal materials, suggesting that the latter resemble collagen in structural and molecular form.

INTRODUCTION Agglutinated foraminifera shells have organic components (linings and cement) of a muco-polysac­charide, or glycoprotein nature (Hedley, 1963), in which the basic biomolecular structure has a pro­teinaceous part linked to an amino sugar/ glycogen ~nit (Stryer, 1975). Further analysis has indicated that the organic material contains, or has deriva­tives of, galactosamine or glucosamine moieties, and hence the term glycosaminoglycans (GAGs) was sug­gested to be more appropriate (Langer, 1992). Additionally the GAGs are thought to be sulphated (Hedley, 1963; Langer, 1992). The presence of sul­phated GAGs provides surfaces which are nega­tively charged and which associate with hydrogen ions. These components readily interact with other charged interfaces i.e., inorganic surfaces, or with positively charged metal ions in solution (Langer, 1992). Details of the precise composition of the organic material are scarce, as many biochemical techniques require substantial amounts of material for analysis and this is not available in aggluti­nated foraminifera (Bowser & Bernhard, 1993).

The application of advanced high resolution spectroscopic techniques in the investigation of organic test materials will enhance previously acquired information. The mechanisms employed by the agglutinated foraminifera using adhesive mate-

rials to aid, and form secure structures, in test wall construction, in aqueous environments, may hold spe­cific information useful in wider applications. Additional information regarding the composition of the organic components (cements and linings) may enhance the established classification scheme. The resistance to degradation, and overall strength of the organic materials, may be important for research in areas of materials science and underwa­ter technology (Bowser & Bernhard, 1993), as strong, and resilient organic materials with good binding capabilities may be advantageous in certain situa­tions. Organically cementing agglutinated foramin­ifera do not biomineralize their test components, differing from other calcareously cementing forms. Hence the organic fraction produced by agglutinated foraminifera must vary from that reported in other biomineralizing organisms (Mann et aI., 1997). Investigation into the form of this more primitive material may reveal basic structural and com­positional differences. Biomimetic investigation (the way organisms produce or sculpt minerals) is a topic of current research interest. The organically cementing agglutinated foraminifera form a gap in the evolutionary process of biomineralization as they have succeeded without the need for secreted minerals, using instead those previously obtained from the surrounding environment.

II!: Hart, M.B., Kaminski, M.A., & Smart, C.W. (eds) 2000. Proceedings of the Fifth International Workshop on Agglutinated Foraminifera. Grzybowski Foundation Special Publication, 7,1-13.

2

The foraminifera used in this study were indi­vidual specimens of Trochammina inflata Montagu, Jadammina macrescel1S (Brady) (temperate marsh foraminifera) and the giant Antarctic foraminifera Astrammilla rara Rhumbler (Plate 1). The foram­inifera selected all have abundant organic compo­nents that can be isolated from the shells with rela­tive ease. Calcareous foraminifera, Ammonia bec­carii (Linne), were used for comparative analysis. Chitin (a glycosaminoglycan) and collagen (a gly­coprotein) (Sigma') were also analysed to provide information as model biomolecular structures.

MATERIAL AND METHODS Samples of Antarctic foraminifera were obtained by methods outlined by Bowser & Bernhard (1993). Temperate marsh foraminifera and calcareous foraminifera were collected from marsh inlets in Swanwick and Warsash near Southampton by ran­dom surface scraping (top few mm of mud). Samples were washed and sieved (63 J1m) and then dried. Specific foraminifera species were picked and stored in separate glass tubes, ready for further analysis.

To obtain detailed information on only the organic part of the shells (lining and interstitial cement), the inorganic mineral particles were ini­tially removed, by techniques typically used in palynology preparations: treatment with hydro­chloric acid (30% w Iv) (2-3 hours), hydrofluoric acid (28% w Iv) (2-3 hours) and then finally boiling in hydrochloric acid (10% w Iv) (20 seconds). Calcareous foraminifera linings were isolated by treatment with dilute hydrochloric acid (10% w Iv). Reference glycoproteins, collagen and chitin, were also treated with similar reagents under the same conditions to see the effect of the chemical treatments.

Analytical methods used were Fourier Transform Infrared (FT-IR) spectroscopy and pyrolysis Gas Chromatography Mass Spectrometry (py-GC/MS) both of which require only small amounts of material for successful analysis.

Infrared spectroscopy Test linings, other organic components (cement) and reference materials were analysed with the FT-IR microprobe. Molecular units, protein and carbohy­drate components, individual amino acids and the protein secondary structure, i.e., alpha helices and beta sheets, can be identified with the use of FT-IR spectroscopy (Haris & Chapman, 1992; Fessenden & Fessenden, 1979; Smith et ai., 1994, Fabian et ai., 1996).

Analysis was performed with a Nicolet™ 460 Protege, Magna IR technology FT -IR microprobe with a Nic-plan™ IR microscope. The data were acquired and manipulated with Omnic software. Infrared bench conditions were as follows through­out the analysiS; acquisition time 200 seconds, band resolution of 4 cm- l

, and gain 2. Background subtrac­tion takes place after each acquisition to remove

K. Allen, S. Roberts & }.W. Murray

additional components from the spectra. The spec­tral range was 4000-600 cm- l

. Noise levels were mon­itored throughout the experiments, and found to be very low.

Foraminiferal components, shells and linings, were additionally treated with deuterium oxide (D20). Samples were placed in glass tubes containing D20 over a period of 24 hours prior to analysis with FT-IR spectroscopy. Crushed foraminifera were used for FT-IR shell analysis. Results obtained from transmitted infrared through a thin layer of shell were of superior quality compared to those from a diffuse reflective mode.

It is often difficult to separate individual sub­components within complicated spectra that strongly overlap and hence the use of methods such as Fourier Self Deconvolution (FSD) have proved to be effective for the separation of strongly overlap­ping bands, increasing the accuracy of measured peak positions (Braiman & Rothschild, 1988). FT-IR spectra with good signal to noise ratios and high wave number precision can be manipulated easily and efficiently by a computer. When the Fourier Transformation is calculated, the width of the resulting spectral band will be less than that of the original convoluted band. Conformational studies of proteins are best carried out in D20, reducing the possibility of band enhancement due to artefacts e.g., the O-H vibrations arising from water molecules. The enhancement technique, Fourier Self­Deconvolution, was applied to all spectra to reveal more information from specific regions. Within the Fourier Self-Deconvolution option within the Omnic computer software enhancement, variable settings for bandwidth and enhancement factors were altered depending on specifiC spectral characteris­tics. Enhancement factors were set below 3, and band-width set to a value approximately <16cm- l

.

Pyrolysis Gas Chromatography I Mass Spectrometry Analytical pyrolytic GCIMS methods are used to obtain fingerprints of tiny quantities of organic materials, which can then be compared to reference traces. Moieties present within the pyrogram, with characteristic mass spectra, can then be potentially assigned to original components which give dis­tinctive pyrolysis products (Van der Kaaden et ai., 1984).

Many papers cover aspects of pyrolysis of sub­stances containing proteins, carbohydrates and lipids (Stankiewicz et ai., 1997; Lapolla et ai., 1992; Briggs et ai., 1995). Foraminifera have organic material within the shells thought to be glycopro­teinaceous in nature. Hence the main pyrolysis products should indicate that there are protein and carbohydrate moieties within the original struc­tures. The Py-GC/MS technique was applied to foraminifera shells, and then to organic material from the foraminifera, T. inflata, J. macrescens, and A. rara. Collagen, a glycoprotein, and chitin, a

Organic components in test walls of agglutinated foraminifera 3

glycosaminoglycan, were used as model reference substances.

To ensure that foraminiferal linings were not con­taminated during preparation procedures, the mate­rial was treated with dichloromethane (DCM) before being placed in a quartz glass tube, which had a plug of silica wool to support the foraminifera linings. DCM was then introduced down the tube with a pipette to ensure that all contaminants were removed. A blank run was performed with only quartz wool, washed with DCM throuph the tube prior to analysis. The quartz tubes (CDS 1000) were placed inside the coil of the pyroprobe and then pyrolysed in a flow of helium (0.65 kg cm3

) for 5 s. The pyrolysis interface was held at 200°C, the GC injector maintained at 210°C, and the flame ionisa­tion detector (FID) at 320°C. The temperature of the GC/MS transfer line was set to 310°C. The GC oven was operated under the following programme: isothermal for 5 minutes at 35°C temperature pro­grammed at 4°C min- l to 310°C, and then isothermal for 10 minutes. The mass spectrometer was operated in full scan mode (m/z 40-650),1 scan/s, 70 eV ioni­sation voltage, 300 mA emission current, and an ioni­sation source temperature of 170°C. Peaks were iden­tified, based on their mass spectral characteristics and GC retention times, by comparison with the National Institute of Standards and Technology (NIST) mass spectral library and published GC and MS data (Stankiewicz et al., 1996). Additionally spectral identifications were based on spectra obtained from reference materials, chitin and collagen.

RESULTS AND DISCUSSION FT-IRi Foraminiferal organic components The main regions of absorption within the mid infrared (600-4000 em-I) for the foraminiferal linings and reference materials occur at 3100-3600, 3100-2800, 1690-1630, 1560-1500, 1400-1200 cm- l and 1180-800 cm- l (Figure 1a-f [1-10D. These regions will be discussed individually and attempts made to relate absorptions to specific molecular vibrations. Treatment of reference glycoproteins with hydro­chloric and hydrofluoric acids under the same condi­tions used for the foraminifera samples did not result in any d_etectable spectroscopic alteration of band frequecies in the residual material.

. At 2956, 2873, 2926, and 2854 em-I, absorptions anse from CH3 and CH2 antisymmetric and symmet­ric stretching modes respectively (Kokot et al., 1994; Bellamy, 1975). These absorptions are clear in all the spectra obtained from foraminifera shells and linings (Figure 1c-f [3D, and are in a region where their identification is not hampered by any p?ten~ial overtone ~ands. These are the only vIbratIons that uneqmvocally identify the presence of organic substances within an agglutinated foraminifera shell.

Region 1 1 J 4 j 6 7 8 9 10

4000 JOOO 2000 1000

Figure 1. Comparison of FTIR spectral absorbances in the 4000-4?Ocm-1 wa,,:e~umbe.r ~ange, in spectra obtained from an.aIY~ls of foranyn!feral h~mgs and standard materials (a­f) mdlc.a!e that SImIlar regIOns of absorption (1-8) relating to speCifIC molecular stretches and bends can be identified. Absorption at 2800-3000 cm-I arise from C-H bonds within organic structur~s (3). Absc:rp~ion at 3088 cm- l relates to CH=CH bond VibratIons wltFun carbohydrate ring struc­tures (2). N-H stretching vibrations, and O-H stretcning vi­brations form a broad absorption from 3000-3700 cm- l (1).

Chitin has well resolved absorptions at 3115 cm- l

arising from CH=CH stretching (Figure 1a [2D, at 2963 and 2876 cm-l (CH)), 2933 cm-l (CH2 ) and 2890 em -1 (CH stretches) (Figure 1a[3D (Focher et al., 1992). Collagen has C-H (stretch) absorptions at 3088 (CH=CH), 2959, 2874 (CH3), 2925, 2853 cm- l

(C H 2 ) these being very similar to those in the foraminifera spectra (Figure 1b[2-3D. The CH=CH vibrations arise from the vibration of double bonds within aromatic ring units of the organic component. Differences in the 3100-2800 cm- l spectral region may reflect variation of C-H bonding arrangements within the materials.

In the organic linings of calcareous foraminifera, a band in the region 1740-1716 cm- l arises from the stretching vibration of a carbonyl (C=O) group of a carboxylic acid unit (Figure 1£[4D (Kokot et al., 1994), probably present in protein side chains consisting of amino acids such as glutamic acid.

The model glycoprotein collagen also exhibits a clear absorption at this wavenumber (Figure 1b[4D. However in the spectra of agglutinated foraminifer­allinings, this band is not observed (Figure 1d,e[4D, although it may be masked by the intense absorption of the amide 1 band at -1660 em-I.

4

In all spectra carbonyl (C=O, stretch) and N-H (bending) vibrations are identified in the 1690-1630 cm'l region, relating to the amide 1 unit, the main peptide link, of a protein (Kokot et ai., 1994) (Figure 1 & 2a-f[5]). Within the amide 1 band, small absorptions can identify features of the secondary structure of the protein i.e., a-helices, b-sheets and turns (Figure 2c-f) (Table 1). The intensities of such peaks may indicate the relative amounts of each type of secondary structure (Smith et ai., 1994). For example, A. rara has a significant absorption at 1635 cm'l (Figure 2c[5]) which may indicate that greater prorortions of beta sheets exist. In. contrast the 1652 cm' absorption is more pronounced In both of the agglutinated marsh species (Figure 2d,e[5]) indicating greater amounts of alpha helices, sug­gesting that J. macrescens and T. inflata have simi­lar proteinaceous secondary structures. With addi­tional enhancement techniques, much clearer band wave numbers and detail can be determined (vide infra). Chitin has definite absorptions at 1660 and 1626 cm'l (Figure 2a), which are assigned to C=O groups hydrogen bonded only to NH groups, and C=O groups bonded with a bifurcated hydrogen both to NH and OH -6 groups, respectively (Focher, 1992). Collagen has a broad amide 1 peak similar to foraminiferal spectra, within which additional absorptions occur relating to the secondary structure of the proteinaceous unit (Figure 2b).

The band at 1577-1505 cm'l and the 1541-1505cm,1 absorption in all spectra (Figs 1 & 2a-f[6]) arise respectively from the O-C-O antisymmetric stretch and the C-N stretch and the N-H in-plane bending absorption in the amide 2 unit of the proteinaceous material (Kokot et al., 1994). The amide 2 region also contains information on the secondary structure of proteinaceous materials as well as detail on indi­vidual amino acids. Vibrations occurring at 1518 and 1498 cm'l (Figure 2<:1[6]) can be detected in all of the spectra obtained from foraminferal linings and are suggested to arise from the amino acids tyrosine, and phenylalanine respectively (Fabian et ai., 1996; Hadden et al., 1995). These specific amino acids are components of the polypeptide chains. With enhancement techniques the vibrations arising from these moieties again become more obvious (vide infra>. Comparing the foraminiferal spectra to those of the reference materials chitin and collagen, it can be deduced that collagen has an amide 2 region which correlates with the foraminiferal amide 2 regions, whereas the absorption in chitin (1563 cm'l) is quite different (Figure 2a[6]). This highlights similarities between collagen and foraminiferal material regarding the amounts of protein secondary components, and perhaps specific amino acids.

At 1470-1380 cm'l band vibrations arising from further C-H deformation, i.e, CH2/CH3 bending modes, occur in all spectra. Between 1400-1380 cm'l the O-C-O symmetric stretches are evident (Figure 3a-f[71). The amide 3, N-H in-plane bending and C-H

K. Allen, S. Roberts & J.W. Murray

a) Chitin

~ {5 c) A.l'tJra

'f o .., 01:)

~

f) CalcareolU IlIIlIIg Iii iii iii I i

11100 1700 1600 1500 W«VIII/lMben (elff'?

Figure 2. Absorption within the spectral regions 4:5 and 6 (a-f), in the 1800-1400 em' I wavenumber ran&e, a~lse from bond vibrations in protein chains. The C=O vlbratlOn, 1750 cm'l , relates to a carboxylic acid bond vibration suggested to be present on protein side chains (Kokot et al., 1994) (4). Carbonyl stretchmg (C=O) and N-H bending vi~rations a~e identified in the amide 1 region (5). O-C-O anttsymmetnc stretches, C-N stretch, and N-H inplane bending vibrations produce absorptions in the amide 2 region of the spectra (6). Within both of these areas small absorptions occur relating to the secondary structures of proteins.

stretches and contributions to O=C-N bending, usu­ally very weak, occur at 1340-1280 cm'l (Kokot et ai., 1994). Astrammina rara has two clear absorption bands at 1450 and 1380 cm'l, whereas a broad absorp­tion at 1380-1400 cm'l is seen in J. macrescens and T. inflata (Figure 3d,e[7]), potentially masking the two individual bands seen in A. rara (Figure 3c[7]). This region could also indicate that there are differences in the type/ abundance of the C-H bonds within the structures. Astrammina rara has similar absorptions to collagen, whereas the absorptions from the other foraminiferal materials are very weak. Chitin and collagen both have absorptions relating to C-H deformation bending modes at -1420 cm'l, and those relating to amide 3 contributions (Figure 3a & b[7]). The S=O vibration of the sulphate group present in sulphated glycoprotein (and similarly phosphate P=O vibrations) may be detected at 1250-1230 cm'l (Sekkal et ai., 1993). As foraminiferal linings are suggested to be sulphated glycoproteins (Langer, 1992), detection in all spectra of absorptions in this region should be possible.

Nor

mal

sp

ectr

a ("

,'av

enum

hers

~ cm

"l)

Fou

rier

sel

f dec

onvo

luti

on (

wav

enu

mb

ers:

em

·l ) S

econ

dar

y st

rucl

ure

s T

roll

illi

er e

t al.

Le G

al e

l QI

. Sk

orko

·e:lI

.lek

Had

den

t't u

l. .'1

.1""1

""

J. "

,a("J"eSCL·II.~

T iI

l1(/

/fl

A.h

ecct

ll"ii

Col

la e

ll

Chi

li"

A.I

"(/n

l J.

lIIa

UC

SC

CII

.,·

r. il

l 1

(1/(

/ A

,hr!

cml"

ji C

olla

'ell

C

hitin

19

96

1991

et

ul.

19

95

19

95

1696

16

95

1696

16

95

turn

, + b

end

, 16

94

beta

16

91

1689

16

89

1690

16

90

1688

tu

rns

+ b

end

s 16

88

1689

1682

16

81

1682

16

82

1683

16

81

turn

s +

bend

s 16

82

1683

16

82

1680

16

73

1675

16

74

1675

16

75

1674

b

ela

shee

t pa

rall

el

1675

+ 4

16

73

1676

16

67

1668

16

68

1668

16

67

1667

16

68

1668

16

68

turn

s an

d be

nds

1670

16

70 +

2

1658

16

60

1661

16

60

1660

16

58

1659

16

60

1660

16

60

1662

16

66

turn

, alp

ha

1660

16

63 +

4

1665

16

52

1652

16

52

1653

16

52

1652

16

51

1653

16

53

1652

15

57

alp

ha

helix

16

53

1654

+ 3

16

56

1657

1645

16

46

1646

16

46

1646

16

45

1644

16

46

1646

16

46

1642

.. I

pha.

ran

dom

eoi

l 16

43

1045

+ 4

16

40

1639

16

39

beta

she

et a

nti

pan

Un

16

37+

3 16

39

1634

16

35

1635

16

35

1635

16

35

1638

16

35

1635

16

35

1635

be

ta s

heet

16

35

1631

+3

1635

16

33

1622

16

27

1628

16

24

1629

16

30

1628

16

28

1628

16

28

bel

a sh

eet

1624

+4

1628

1623

16

25

1623

16

22

1624

16

20

beta

she

et

1625

16

15

1616

16

16

1616

16

18

1616

16

16

1616

16

12

tyro

sine

16

12

1616

16

07

1608

16

09

161

I 16

08

1608

am

ino

acid

sid

e ch

ain

1605

1567

15

68

1569

15

65

1563

15

63

1570

15

71

1569

15

70

1568

"O~;3"';

0 ag

-o.:

~ ... CICl

ro a

I ~

'" S'&~'~

:I

,,~::lt"'

.... o

i'1

@:j

,., ,.,

:3 ('

0

1-1

,...I

0 "0

C/

l It>

It>

;3

° ~ o..~

::l1

il",

0"

"0

~ c.1

1 Q

0 :I

ore~cJ

III

~rti ~

~

:I ....

~~

1-1--.

~ C

/l

~ ~a ;2

s·

~::l

;3~

.... III

~

O"

......

1t>

C/l

2.;2

::;

::l ....

" ~ ~ ~

~

s-;;;

;3 ;3

a S'

::l 5'~

-C/l ()

q @

::;.;~

0

;3 It>~

.....

0","

,,,

'" ~ ~ e

.;3

CICl

~

V"l

a .!.

CI

Cl I-'.

(j 1

-1

_

-e:

,.:.

loq

("")

=

.

::l ~

::l"

::t.

O"::l~

:I

('0

-'1

-1

" ~

'" ~ ~"

.... ti

l _

• ...

,. II

I ;Q

·8.~

Q

.. cg

::l tn·

'"" 0

a~ ~

. ... '"

1564

15

63

1564

15

65

1564

15

65

1563

15

62

beta

,h

eet

1564

15

58

1558

15

58

1558

15

58

1558

15

57

1558

15

58

1558

15

54

1555

15

50

1552

IS

S I

1550

15

51

alp

ha

rand

om

1550

15

48

1548

15

49

1548

15

49

1548

15

45

1549

15

48

1547

15

43

alph

a he

lix

1545

15

39

1539

15

41

1540

15

43

1540

15

38

1540

15

40

1543

8">

0'

;3 .....

. ..,

.... C/

l ~"O

:I

11 ~ a

.... '""

§.r§

N. II

I

""0

::l

~ 15

32

1533

15

35

1539

15

33

1530

15

34

1533

15

39

1535

be

ta s

heet

15

32

1532

15

25

1527

15

23

1527

15

27

1527

15

27

1524

15

22

1520

15

20

1520

15

17

1522

15

20

1520

ty

rosi

ne

1518

1.

116

1519

15

16

(Jj~(Jj

It>

" It>

"

~ "

g 0"

. °

o..o

::l

~::lfr

.., ° ..,

'< .

..... '<

15

16

1516

15

1l

1513

15

16

1514

15

13

1512

15

09

1506

15

07

1508

15

06

1508

15

06

1506

15

05

1504

14

98

1499

14

95

1492

14

95

1492

14

92

1491

14

87

1490

14

93

1487

14

85

1485

14

83

1481

14

84

1479

14

78

phenyl~lanine

1471

J

1474

14

71

1470

14

75

q. ~~

~

::: :::

" ..,

" 2"

~ ...

. ~

~ E;

til v:d~

~~5·

14(,

) 14

63

1463

14

66

, (t

og

. rti

('0

('

0

::l "

~ f6

g ;3

til

<:: -.

,,

00

..

_ .

..-f

'D

rts.~

p.. -.

QJ

s· g

5. ~"'N

~ff~

1

t>C

/l()

q

-"'0

-.

° It>

°

...... ,

,::l

0"

"1" _.

tIl

::r'::

::!':'o

(J

1

It>

"

......

6

Region 7 8 9 10

a~~~~ ____________________ ~~~

f) Calcareous lining

1400 1100

'"

1000

Figure 3. C-H bending deformations, O-C-O sym~etrical stretches and amide 3 bonds produce weak absorptions In

all spectra (a-f) from 1300-1450 cm'l (7). Absorptions at 1250 cm'l in foraminifera spectra are assignable to 5=0 bond vibrations suggesting that the materials may be sul­phated (c-f 8). Characteristic absorptions fro~ CoO ~o~d stretching vibration of carbohydrate umts within biomolecular structures occurs at 1200-1000 em'! (9).

The absorption at 1250 em'! of varying intensity in the spectra of T. inflata and J. macrescens (Figure 3d,e[8]) may be attributed to the presence of sulphate. Astrammina rara (Figure 3c[8]) has an absorption at -1250 cm'l in some, but not all spectra, which may suggest that the component may not always be present, lacking in either sulphate or possibly phosphate groups. !" w~a.k absorption at 1250 cm' can be observed m chitIn and collagen spectra (Figure 3a,b[8]). These structures are not sulphated, and hence this absorption may arise from other skeletal vibrations. Thus the 1250 cm'! band absorptions in the spectra of the foraminifera A. rara, J. macrescens, T. inflata and A. beccarii rna y also arise from vibrations of other bonds within the molecular structure. The presence of sulphate ions was not detected after hydrolysis with boiling hydrochloric acid and latter addition of aqueous barium chloride (BaCI2). However, if the organic structures are not destroyed by mild acid hydrolysis, then the sulphate ions may not have been detached from the structures.

The region 1180-1000 cm'! is a very characteristic area where CoO stretches arising from carbohy­drate/polysaccharide units occur (Sekkal et aI.,

K. Allen, S. Roberts & J.W. Murray

1993). This region of absorption within the four species analysed has the most variation (Figure 3a­f[9]). In A. rara, peaks of similar intensity occur at 1182, 1108, 1059, and 1001 cm! (Figure 3c[9]). The marsh foraminifera, T. inflata and]. macrescens spectra also have peaks at 1062 cm'! (similar to 1059 cm'!) and additionally two weak absorptions at 1164 and 1122 cm'! (Figure 3d,e[9]). This could indicate variations in the nature of polysaccharide compo­nents. The absorptions in the spectra of calcareous foraminiferal linings vary quite significantly. An intense absorption at 1027 cm'! in the latter does not correlate with any of the absorptions within spectra from the other agglutinated foraminifera, although it is potentially similar to the strong 1025 cm'l absorption visible in the chitin spectra (Figure 3a, f[9]). Weak absorptions at 1056 and 1152 cm'! in the spectra of calcareous foraminifera are similar to those in the spectra of the other agglutinated foraminifera (A. rara 1059 cm'!, T. illflata and J. macrescells 1062 cm'l) (Figure 3d,e,f[9]). Comparison of the same regions in the spectra of the collagen and chitin reference materials with those of the foraminiferal linings show significant differences. In collagen and chitin, absorptions at 1166 cm'! and 1157 cm'! may relate to similar absorptions within the marsh linings (Figure 3a,b,d,e[9]) (agglutinated and calcareous). Additionally the 1118 cm'! band in chitin, and the 1117 cm'l absorption in collagen may arise from similar vibrations to those of the weak 1122 cm'! absorption identified in marsh foramini­feral linings (Figure 3b,d,e[9]). There is significant similarity between the 1060 cm'l absorption present in the chitin spectra, attributed to the CoO stretch of the N-acetylglucosaminoglycan residue (Pierce & Rast, 1995) and the peaks of approximately the same wavenumber in spectra from all foraminiferal materials (Figure 3a,c-f[9]). This could suggest that similar carbohydrate ring formations exist. The spe-

. ,J cific assignment of the bands In the 1200-1000 cm region is quite problematic as a number of strong absorbances occur, arising from various groups.

A clear absorption at 826 cm'! and a shoulder at 1000 cm'l are present in the spectra of A. rara and J. macrescens (Figure 3c,e[10]). This is suggested to be similar to the vibration from the equatorial configu­ration of the -OS03H groups of the carbohydrate component observed in a study of sulphated glyco­proteins in human saliva (Embery et aI., 1989). The strong absorptions between 1230-1250 cm'! arising from to the S=O stretching (of the ester sulphate groups) are additionally discussed in their analysis. Hence the -1250 cm'! absorption in the foraminifer­al spectra may be attributed to the presence of sulphate within the organic materials, correlating with the 826 cm'! band.

The broad absorption at 3100-3600 cm'! within all four foraminifera lining spectra provides additional evidence that both proteins and carbohydrate com­ponents exist within foraminiferal lining materials (Figure 1a-£[1]). N-H bond stretches within a pro-

Organic components in test walls of agglutinated foraminifera 7

:.:inaceous components (Kokot ct (Ii., 1994), vibra­::ions arising from O-H bond stretches within carbo­:-:\-drate units and also as water held within the ~~ructure will form absorptions in this region.

Deuteration of foraminifera components \',nen organic material is treated with deuterium (,,,ide, any easily exchangeable hydrogen atoms are replaced by the deuterium of heavy water. The O-H and the N-H bonds have molecular vibrations that [.:sult in an intense absorption at -3500 cm· l

. After lieuteration of the reference materials, linings and ~hdls, an absorption can be seen at 2700-2400 cm·1

,

and the 3300-3600 cm- l absorption decreases in intensity (Figure 4a-f[1,2]).

Exchange of hydrogen with deuterium in O-H and ='J-H bonds would produce the observed features. The amide 2 regions are altered on deuteration. The ::'and at -1550 cm- l decreases in intensity, which relates to exchange of N-H to N-D, and shifts 100 cm- l towards the 1450 cm· l region, adding to the band iue to C-H deformation bending modes. The absorp­:ion remaining in the region of 1550 cm-l is then due onl\- to the C-N stretch (Fabian ct ai., 1996). The alt;ration of the amide 2 band is observed in all of :he lining spectra (Figure 4a-e[3,4]). The amide 1 C=O band is not affected. The apparent change in the amide 2 band is of great value in monitoring H-D .:xchange, and relates to changes in protein confor­:l1ation due to ligand binding, pH and temperature Haris & Chapman, 1992). The 1000-1300 cm- l region

in the foraminifera spectra becomes slightly altered upon deuteration although the main differences are \,'ithin the intensities of the peaks and not the actual wave number (Figure 4a-f[5]). In the deutera­hon experiment on the calcareous foraminiferal lin­;ng, the bands from 1200-1000 cm-l decrease in inten­~it\', broaden and are less well resolved. The effect ot"D20 exchange on collagen was quite significant. The weak absorptions at 1300-1340 cm- l increased in intensity on deuteration to form a strong absorption at -1340 cm- l probably masking the lower band at 1300 cm- l (Figure 4b). A slight decrease in absorption due to the amide 2 band, after exchange of N-H for ~-D bonds, was apparent with the expected increase in intensity in the 1450 cm· l absorption I Figure 4b[3,4]). This may reflect modification of O=C-N bending modes. In deuterated chitin, absorp­tions from the amide 1 and 2 bands are only slightly lower in intensity, whereas the bands at 1430, 1330, 1250, 1200, 980, 950 and 900 cm- l decrease in absorp­tion quite significantly (Figure 4a[3,4,5]).

This study supports Langer's view (1992) that foraminiferal organic components are glyco-samino­glycans. The broad effects, changes in both the amide and the carbohydrate regions, observed in the collagen standard are also apparent in the foraminifera spectra, although the variations in original absorption wavenumbers indicate that the structures are not identical i.e., there are slight \'ariations from one species to another.

1 J4 j

4000 3000 2000 Wav~"UIfIh~T.I (t!Ift'~

Figure 4. Comparison of regions (1-5) in all spectra (a-f) in the 4500-450 cm-l wave number range after D20 treatment with original spectra indicates that the absorption ~t 3000-3700 cm-l is lowered due to H exchange wlth D 1Jl O-H bonds (a-f 1) with addition of absorptions at 2500 cm- l (a-f 2). A decreas~ of the amide 2 absorption due to replacement of N-H by N-D (100 cm-l shift) causes an in~r~ase of the1450 cm-l C-H bending vibrating (a-f 3~4). Ac!.dltlonally absorp­tion frequencies in the fingerpnnt regIOn of an spectra appear to become weaker (a-£ 5).

Fourier self deconvolution (FSD) The application of FSD to the spectra of foraminiferal linings, and also those treated with deuterium oxide, enables greater resolution of bands relating to those of protein secondary structure, in the amide 1 and 2 regions. In most cases, from FSD traces, regions involving protein sub-structures, e.g., alpha (a) helices, beta (b) pleated sheets and turns and specific amino acids were determined. All of the foraminiferal linings within the IR study have the same proteinaceous secondary structures (Table 1), although initial spectra (not enhanced) indicate that the secondary structures are present in varying amounts (Figure 2c-f [4-5]).

The enhancement technique was also applied specifically to tests of foraminifera which con­tained inorganic components (quartz). Vibrations of the inorganic component may mask vibrations assigned to the amide 1 and 2 regions, although results from FSD indicated that the same bands which were assigned to secondary structures in the linings could also be assigned in the shell analysis (Table 1, Figure 5a,b[2]).

8

1700 1600 W~lIll1Nben (CIII-,)

1500

Figure 5_ Comparison of spectra a and b in the 1500-1800 cm-) wavenumber range indicate that the technique Fourier Self Deconvolution enhances absorptions in the amide 1 and 2 spectral regions clearly defining the absorptions characteristic for protein secondary structure in lining and shell spectra.

After the application of FSD, to both lining and shell spectra, absorbances within these regions are clearly distinguishable and relate to specific wavenumbers associated with the presence of alpha helices, beta sheets, turns and bends, and addition­ally individual amino acids. Specific information about organic components, cement and linings, can therefore be obtained from simple IR probing of shell material. The detailed structure of the organic components of foraminiferal shells with the FT-IR FSD enhancement technique, reveals that the organic material within the foraminifera shell analysed is very similar in composition to the lining material (Figure 5a,b[2]). FSD was also applied to deuterated foraminiferal components. The results, when compared with original spectra, displayed the same wavenumbers in the amide 1 region and amide 2 region of the infrared spectra.

Metal ion interactions with foraminifera shells Foraminifera shells are often discussed as having iron-rich components, presumably Fe3

+, as the shells are often brown/orange in colouration. Iron was found to be present in the brown/orange regions by SEM techniques (Allen et al., unpublished work). Occasionally the last chamber to be formed remains white, potentially without oxidised iron in or on the surfaces of the chamber. Whether unoxidised Fe2

+ is present is unknown, but a colour difference is distinguishable. FT-IR can characterise differences within the areas of shell wall leading to a poten-

K. Allen, S. Roberts & J.W. Murray

tial hypothesis about how the iron interacts with the organic or inorganic components of the shell. Foraminifera with final/new white chambers, and older iron-rich chambers were dissected, and anal­ysed over the 4000-450 cm-) wavenumber region. The spectra from the iron-rich chambers were very simi­lar to those from the white chambers, except for the presence of a new band at 1038 cm') and the absence of an absorption at 830 cm-) in the iron rich spectra (Figure 6a) In foraminifera shell analysis, the amide 1 and 2 regions (1660-1550 cm'\ appeared not to be affected by the presence of iron (III) species (Figure 6a,b), This region was affected in wool keratin which was treated with high metal ion concentra tions (Kokot et al., 1994), Noticeable differences were not evident in the amide regions in spectra of the iron rich chambers.

a) Iron rich chamber b) Final chamber c) Quartz

2000 1500 Wal'tlllllllbers (CIII")

iii

.1000

Figure 6. Comparison of spectra obtained from an iron rich chamber, a non iron coated specimen and quartz (a,b,c). The lack of the 830 cm') absorption in the iron rich spectrum, in a complex region of Si-O bond vibrations, suggests that iron may be interacting with inorganic surfaces in quartz enhancing Si-O vibrations in the spectra (iii). The amide 1 and 2 regions appear to be unaffected by the presence of iron (i,ii).

As the vibrations from the foraminiferal analy­sis lie in the complex fingerprint region of the spec­tra the iron could be interacting with inorganic material, or amino-acid residues (potentially cys­teinal residues with high affinities to iron) of organic components (Kokot et a/., 1994), In the iron (Fe-III) rich shell spectra the lack of a 830 cm') absorption is apparent within a region of molecular absorptions arising from Si-O bond vibrations (Figure 6a[iii]). It is suggested that Fe is binding with inorganic minerals present within the shell rather than the organiC materiaL Quartz surfaces have high affinities for metal ions (Lo et al., 1997). The interaction of metal ions with inorganic rather than protein moieties may be suggested in view of the lack of change in the intensities of the amide regions (Kokot et al., 1994; Nahar et az', 1995) and also the lack of detail within the carbohydrate areas of the spectra due to the overtones of Si-O bond vibrations. This would also assist in discussion of how the iron phase interacts with the

Organic components in test walls of agglutinated foraminifera 9

:.'c'::aminiferal wall components. The suggestion that :r.e incorporation of iron is due to secondary mecha­:-:b-ms of chelation onto quartz surfaces rather than ?roduction within the foraminiferal internal pro­.::esses agrees with the view of Sidner et al., (1976). iron within the surrounding environment would ?Otentially react with both organic and inorganic s:rrfaces. Iron in the shells can be detected with the Electron X-ray probe (ED AX) of the Scanning Elec­:ron Microscope, and also sulphur and phosphorus can be identified in very small quantities. FT-IR analysis displayed characteristic absorptions which may arise from the vibrations of S==O, and OS03H groups within the organic structures. The 5Ulphate containing groups are likely to form nega­::Yely charged surfaces within the biomolecule, and b.ence there is no reason to exclude pOSSibility of additional sulphate group - iron interaction.

Pyrolysis GCMS; Organic components Products visible within the pyrogram vary accord­:ng to their retention times within the chromato­graphic column. Simple moieties, of low molecular mass, are liberated initially. All the spectra have extremely large initial peaks that are due to the volatilisation of excess DCM, and other easily yolatilised substances. Thereafter, peaks assigned to acetic acid, and other substances with m/z 44-60 are apparent. These initial peaks may derive from pyrolysis of either carbohydrate or protein units and generally are not very useful for specific analysis/identification (Figure 7a, Table 2a).

Pyrolysis products forming after these initial peaks are more important for revealing information about the complex organic substances present. Pyrolysis products include pyridine (m/z 52,79), pyrrole (m/z 39,67), toluene (m/z 51,65,91), phenol ·m/z 39,66,94) and styrene (m/z 51,78,104). 2-~fethylpyrimidine (m/ z 41,66,93), 4-methylphenol (m/z 39,51,77,108) and 1H-(2 methyl) pyrrole (m/z 39,66,93) are also abundant. Indoles can be identified (m/z 90,117), together with methylated 1H-indoles (m/z 42,51,44,133). Designation of these pyrolysis products is relatively straightforward (Figure 7a, Table 2).

Protein indicators - Phenols and methylphenols are pyrolysis products from the amino acid tyrosine (Stankiewicz et al., 1996) although the observation of abundant phenol and methyl phenols may not be specific for tyrosine since these products may also originate from a non-proteinaceous phenolic bio­polymer (Briggs et al., 1995). The formation of pyrrole indicates that hydroxyproline and proline are present within the original biomolecule (Stankiewicz et al., 1996; Briggs et al., 1995). Toluene, styrene and cyanobenzenes are indicators of p/zenylalanine (Stankiewicz et ai., 1996; Briggs et ai., 1995). Ethylbenzene, phenylacetonitrile, and phenyl-propanonitrile, additionally arise from phenylalanine (Stankiewicz et al., 1996). Indole

and methyl indole derive from tryptophan (Stankiewicz et al., 1996). In pyrolysate analysis of collagen, residues of phenylalanine and tryptophan are found only in trace amounts or are often absent (Briggs et al., 1995). Earlier studies of the composi­tion of the organic material in foraminifera have suggested that collagen-like structures are present, and hence only trace amounts of the latter pyrolysis products would be expected in GCMS-pyrolysis experiments, as has now been found. Other pyrolysis products identified in A. rara, J. macrescens, T. inflata, and collagen and chitin reference materials (Sigma') were diketodipyrroles (m/ z 43,65,93,130, 158,186), and diketopiperazines (DKP) (m/z 70,154,168,194) (Figure 7, Table 2). Diketodipyrroles were identified in the collagen pyrolysate and probably arise from pyridine, or other amino carbo­hydrate units (Stankiewicz et al., 1996). Diketopiperazine moieties derive from adjacent amino acid pairs (Mauger, 1971), in a double dehy­dration reaction of amino acids yielding dipeptide and the diketopiperazine system (Ratcliff et al., 1973). DKP's form from the pep tides involving pro­line, alanine, arginine and glycine as predominant amino acid moieties (Stankiewicz et al., 1996). One of the designated pyrolysis products from A. rara was identified as a proline-proline (m/z 154) 2,5-diketopiperazine. Similar products are found within the chitin py-GC/MS analysis (Stankiewicz et al., 1996), indicating that proline is identifiable within the foraminiferal macromolecules. Caprolactam, identifiable in all traces, is often found in the pyrolysates of nylon (Figure 7, Table 2[26]). Potentially this pyrolysate may be a contamination from the sieving process after acid treatment, or perhaps introduced in the initial stages of processing when samples were in synthetic vials. Storage of the organic components, after sieving, in methylated spirits, and then treatment with DCM all within glass tubes should have removed the effects of contamination, and hence the origin of caprolactam appears problematic.

Carbohydrate indicators - The elutants 2,3-dihy­drobenzofuran and 2, 4 - dimethylfuran are found in the pyrograms of all the foraminiferal organic materials in varying intensities (Figure 7, Table 2a[11&25]). Previous py-GCMS studies of polysac­charides indicate that furans are the predominant pyrolysis products (Bracewell et al., 1989). Hence the origin of the furans is suggested to be related to the presence of carbohydrate moieties within the organic materials. This may verify that the linings are glycoproteinaceous. In a Py-GC/MS investiga­tion of conifer seed and cone (Stankiewicz et al., 1997) polysaccharides were identified by the pres­ence of 2-furaldehyde, 5-methyl-2-furaldehyde, 2-furanmethanol, 2-cyclopenten-2-one and its methyl homologue. Additionally 5-methyl-2-furanone and 1,6-anhydro-f5-D-glucopyranose (levoglucosan-40) were identified, together with anhydrosugars, e.g.

10

aj Jadammina "'acrescens - O'1:lnl< llnlne

100

K. Allen, S. Roberts & J.W. Murray

£+05 2.014

1000 2000 JOOO Retention tl_ (secollllsj

4000 5000

b j Jad4.",mintl ....u:resans - .htU

10

1000 1000

£+05 2.000

5000

Figure 7. Comparison of components of py-GC/MS pyrograms (a and b) after analysis of linings and shells of J. macrescens suggest that protein is the main component of the organic material in both cases, with additional minor amounts of interstitial caroohydrate and also lipid components (letters relate to Table 2).

l,6-anhydro-B-D-glucofuranose, eluting between C2 -

phenols and guaiacylpropan-2-one (Stankiewicz et al., 1997). The use of py-GC/MS to investigate the interaction of glucose with amino groups of proteins indicated that furan and pyrrole moieties often arise following glycolation (Lapolla et al., 1992). Many of these complex products are not present within the pyrolysis products from the foramini­feral analysis, which might indicate that the levels of carbohydrate moieties within the general structure are low. The main pyrolysis components indica te that the grea ter part of the rna terial appears to be proteinaceous in nature.

Additional components - Long chain hydrocarbons and their derivatives were detected in all pyrograms from foraminiferal linings (Figure 7, Table 2a[30-32]). These fragments usually arise from substances known to have a lipid content. In the analysis of whole organisms, the presence of alkenes and alkanes reflects the presence of fats and waxes (Simmonds, 1970). Alkanes and alkenes are com­monly found in pyrolysates of virtually all types of kerogens and are characteristic products of resistant aliphatic biopolymers that occur within algal cell walls and higher plants (Tegelaar et al., 1989). The alkanes and alkenes may therefore arise from the presence of lipids within the cell wall of the foraminifera. Chemical analysis of organic mate­rial within the shells of A. rara, identified lipid

material, suggested to contribute to -19% of the overall composition (Delaca, 1986), although this was suggested later to arise from pseudopodial con­tamination in their analysis (Bowser & Bernhard, 1993). More recent analysis of the organic content of A. rara suggested that protein was an integral com­ponent of the bioadhesive, but excluded mention of carbohydrate or lipid moieties (Bowser & Bernhard, 1993). In the case of the organic compo­nents, the presence of lipids has been questioned, although in pyrolysis of organic material, fatty acids are eluted within all of the pyrograms. Measures were taken to remove the effects of con­tamination and hence such pyrolysiS products do appear to arise from the foraminiferal organic material, although this has to be subject to some uncertainty. In pyrolysis experiments on glycated albumin, esters and fatty acids were eluted although found to be contaminants as they are not pyrolysis products of proteins (Lapolla et al., 1992). The results of pyrolysis GC/MS analysis on foraminiferal linings indicate that proteinaceous pyrolysates are far more abundant than those from carbohydrate units. It is possible that preparation techniques may have removed some of the more eas­ily hydrolysed carbohydrates, leading to the con­clusion that the predominant component remaining is protein with minor amounts of carbohydrate, but potentially still a glycoprotein. The analysis of shells from the agglutinated marsh foraminifera

Organic components in test walls of agglutinated foraminifera 11

:rab1e 2. List of mal'or products in the pyrolyzates from foraminiferal linings and shells of J. 111acrescens. Numbers al-32, and :'1-21 refer to pyrOlysis products within pyrograms a and b in Figure 8. Mass numbers (m/z) are based on interpretation of ::taSS .spectra and MIST mass library, and after Stankiewicz et al. (1996). Suggested origins for pyrolysis products are based 0;: intormation from literature reviews (see Stankiewicz et al., 1996).

a) J. macrescens - lining Retention time l\1;lSS numbers m/7.

240 45.60 411 52.79 433 39.67 4/4 o .' "Uc JY."-bb.<.:.", b," .I' .".<U.< bYI 4U.5J,<U.<

730 39.65.91.106 757 39.44.77. 91.106

1.] 805 51.78.104 11 820 39.67.96.106

850 39.45.56.69.83.97 900 39.53.67.80.94

!4 910 38.60.84.94 930 3951.65.80.94. I 07

i6 990 91,120 UY. JY,bb,94 IMI

4lJ. 6. _0 IL<U

1320 42.58.70.127 1327 39.53.77.107 1404 39.51.59.90.117 1645 91.131 16RO 39.91.120 1714 42.55.69.84.1 13.121 1810 90.117

'8 2184 42.51,104,133 2%0 Iu.154 .I IIY 4.1.60, J,8.,,"I, 115, 12" JJn J411

b) J. macrescens - shell ~umOcT Retention time I\'lass numbers (m/z

140 45.60 325 123 375 52.79 391 71

" ""0 jV,OI

lIZ

990 103 10 1040 39.66.94 II 1232 39.77.108 12 1286 39,77.108 13 1375 39.5\.59.90,117 14 1515 122 15 1624 145 16 1635 120 17 1782 90,117 .. LU LOLL

L. ,.L.

should be able to help with this issue since if the preparation techniques remove more readily soluble components, then these should be present in the ini­tial pyrograms from the shells. The origin of the long chain hydrocarbons is suggested to arise from a lipid component within the foraminiferal organic material.

Agglutinated shell analysis Although pyrograms were weaker within the shell analysis the spectra indicate that the main components are those coming again from protein­aceous materials, with the pyrolysis products being identified as phenol, 4-methylphenol, toluene, pyrrole, styrene, cyanobenzene, and indole (Figure 7, Table 2b). Evidence of furans, indicating the presence of carbohydrate components, appears to be weak, indicating that the preparation procedures did not remove carbohydrate moieties from the

Name Origin Ul'l:tic acid

DHidine aminosu 'ar? pyrro!e proline

10 Ilene r lcny a anlne p) nOme -met 1y UlnlllosLlgar .

-pyrro e, -met 1y pro Inemyaroxypro Ine

-pyrro e,_-methy pro mel lyaroxypro me

ethyl bcn7cnc ,ht.:nvlalanine ethv I benzene

styrene 'lhenylalanine ).4 dimelhvlfuran carbohydmte

cyclopropane \ I [ vITale. 2.3 dimelhvl hydroxy rolinc

1 H f'\\[Toic. 2,4.dimcthvl hvdroxYoroline III pvrro\e. 2.S.dimethvl hvdroxvoroline

benzene.propyl phenylalanine

plcno tyroslile en7ene, ell)' Illt:t ly pheny a aml1t:

pyrro I me. mel) phcno . 4 mell) tyrosine

2-4. Illlidazolidiene-dione.3.S.5 Irilllelhvi Dheno! 3-lllclh I tyrosine

henvlaCdonitrile phenvlalanine phCllvl1r() )anonitrilc henvialanine

benzofuran.2.3 dih\dro ea lrolactam

Indole Il"\ntolhan

1 H indole 2,3 dihydroxo-4 methyl tryptophan

I "etoplpenZine pru me-pro me

l.:xadccJJlOIC aCid IPld

14-pent.1 ccanOIC aCI IpU.l

OCl3dccaniC aCid Ipld

Name Ori in acetic acid

nitrobenzene phen lalanine pyridinc aminosu!.',ar?

pyrrolidinc pm:ny a anlnc

p)-fro c pro me

!>lyrene p lell~ a Ullllle

~.q GlIllet I) uran CJmony rate

c)"anO/x111Cne )hen'y la Inn int: henol I 'rosine

2. mcth ilphcnol t'rosinc 4. mClbYlJlhcnol lyrosine

~hcnvlncctonitrilc phcnylalanine 4. elh lphcnol tyrosine

3-(4 methylphenyl) ropanoniIriic pllen lalanine 4-vin I henol tyrosine

indole tryptophan

.::-mellOXY. -vln)' pleno yrOSll1C

U Inuoe tryptopllan

OIKCloplpcrlZlnc cxanoecanOIC aClo Iplu

original macromolecular structure (Figure 7, Table 2b[8]). Evidence for long chain hydrocarbons can also be found within the pyrogram, indicating the presence of lipid components (Figure 7, Table 2b[21]). The presence of inorganic material is thought to interfere with the pyrolysis process, resulting in weak pyrograms. The removal of calcareous components from cuticles in certain organisms was found to enhance the resulting pyrograms quite significantly in chitin analysis (Stankiewicz et al., 1996). Hence the pyrolysis of only the organic material is thought to be a more reliable technique.

CONCLUSIONS Results from FT-IR and Py-GC/MS analysis of the organic materials in agglutinated foraminifera tests have highlighted the presence of protein and carbohydrate components, thus supporting earlier suggestions that the cement and lining material is

12

glycoproteinaceous. The most intense regions of absorption in FT-IR spectra and additionally the predominant pyrolysis products within py-GCMS analysis indicated that the main part of the organic material exists as a proteinaceous component, con­taining only minor amounts of carbohydrate. The secondary structure of the protein component has been described for the first time indicating that variations exist within the organic components from Antarctic and marsh agglutinated foraminifera. Comparison of glycoprotein reference standards with foraminiferal materials suggests that the organic materials within the tests are similar to collagen rather than chitin.

The basic biosynthetic pathways are similar in all eukaryotic cells including the foraminifera. Ribosomes secrete the initial proteinaceous materi­als into the endoplasmic reticulum and then are fur­ther modified by the Golgi apparatus (Whaley, 1975, Langer, 1992).

Additional chemical analysis for exact determi­nation of carbohydrate units needs to be performed. However, a major difficulty is that the lining mate­rial is extremely resistant to acid hydrolysis and hence individual structural units that are required to be detected are not readily available. Thus tech­niques applied to the study of organic material within agglutinated foraminifera where whole structures can be investigated are the only practical possibilities, although not quite as revealing. Further work may indicate exactly how carbohy­drate units interact with the proteinaceous parts of the biomolecular structures.

The presence of sulphated components may be deduced from FT-IR analysis, in agreement with earlier suggestions that sulphated glycoproteins exist within the agglutinated foraminifera organic test materials (Langer, 1992). The sulphur may be present as a sulphated carbohydrate unit or addi­tionally as cysteinal amino acid side chains. The presence of sulphated moieties would increase pos­sible metal interaction with proteinaceous materi­als, although surfaces of inorganic materials are extremely susceptible to metal chelation.

The observation of lipid pyrolysis products causes complication as measures were taken to remove possible contamination and hence these components appear to arise from the organic mate­rial within agglutinated foraminifera, although this is still subject to some uncertainty.

FT-IR and py-GCMS analytical techniques may further prove to be useful techniques for comparison of organic materials within tests of agglutinated foraminifera from varying environments to identify structural and potentially compositional variations.

ACKNOWLEDGEMENTS This research is funded by a Southampton University scholarship. Specimens of Astrammina rara were obtained by S.S. Bowser through grant 92-20146 from the National Science Foundation, Office

K. Allen, S. Roberts & J.W. Murray

of Polar Programs. Professor D.W. Allen is thanked for constructive discussion on FTIR spectroscopy. The organic geochemistry unit, Bristol university are thanked for the use of the Py-GCMS facilities sup­ported by NERC, and helpful discussions with Dr. A. Stankiewicz, Dr. P. van Bergman and addition­ally Jim Carter and Andy Gledhill are thanked for supervision on equipment.

REFERENCES Bellamy, L.J. 1975. The Infrared Spectra of Complex

MoleCllles. John Wiley and Son. New York. 433pp. Bowser, S.s. & Bernhard, J.M. 1993. Structure, bioadhesive

distribution and elastic properties of the agglutinated foraminifera test of Astrammina rara (Protozoa; Foraminiferida). Journal of Eukaryotic Microbiology, 40,2, 121-131.

Bracewell, R.N. 1989. The Fourier Transform. Scientific American, 260, 6, 86.

Braiman, M.S. & Rothschild, K.J. 1988. Fourier transform infrared techniques for probing membrane structure. Annual Reviews of Biophysics and Biophysical Chemistry, 17, 541-570.

Briggs, D.E.G., Kear, A.J., Baas, M. de Leeuw, J.W. & Rigby, S. 1995. Decay and composition of the hemichordate Rhabdopleura: implications for the taphonomy of graptolites. Lethaia, 28, 1, 15-23.

Delaca, T.E. 1986. The morphologX and ecology of Astrammina rara. Journal of Foraminiferal Research, 16, 3, 216-223.

Embery, G., Green, D.RJ. & Rolla, G. 1989. Structural probe analysis on the attachment of Salivary Glycoprotems to Hydroxyapatite Using Fourier-Transformed Infrared Spectroscopy. Caries Research, 23, 247-25l.

Fabian, H., Yuan, T., Vogel, H.J. & Mantsch, H.H.1996. Comparative analysis of the amino and carbonyl terminal domains of calmodulin by FTIR spectroscopy. European Biophysical Journal, 24, 195-201.

Fessenden, RJ. & Fessenden, J.S. 1979. Organic chemistry. Wadsworth International Student Edition. Wise. 1122pp.

Focher, B. 1992. Chitosans from Euphausia superba. 2: Characterization of solid state structure. Carbohydrate Polymers, 18, 43-49.

Green, D.R.J. & Embery, G. 1985. Structural studies on a sulphated glycoprotem preparation isolated from human salIva. Archzves of Oral Biology, 30, 859-86l.

Hadden, J.M. Chapman, D. & Lee, D.C 1995. A comparison of infrared spectra of proteins in solution and crystalline forms. Biochimica et 13iophysica Acta, 1248, 115-122.

Haris, P.I. & Chapman, D. 1992. Does FT-IR spectroscopy provide useful information on protein structures. TIBS, 17,328-333.

Hedley, RH. 1963. Cement and iron in the arenaceous foraminifera. Micropaleontology, 9, 4, 433-414.

Kokot, S., Cheng, J. & Gill, N. 1994. Comparative study of metal ion interactions with wool keratin using chemomet­rics. Analyst, 119, 677-681.

Lapolla, A., Gerhardinger, c., Baldo, L., Crepaldi, G., Fedele, D., Favretto, D, Seraglia, R, Curcuruto, 0., & Traldi, P. 1992. Pyrolysis - GC/MS in the characterisation of gly­cated albumin. Journal of Analytical and Applied pyrolysis, 24, 87-103.

Langer, M.R 1992. Biosynthesis of Glycosaminoglycans in foraminifera: a review. Marine Micropaleontology, 19, 245-255.

LeGal, J.M., Manfait, M. & Theophanides, T. 1991. Applications of FTIR spectroscopy in structural studies of cells and bacteria. Journal of Molecular Structure 242, 397-407.

Lo, S.L., Jeng, H.T. & Lai, C.H. 1997. Characteristics and adsorption properties of iron-coated sand. Water Science and Technology, 35, 7, 63-70.

Organic components in test waJIs of agglutinated foraminifera 13

\lauger: A.B. 1971. Degradation of peptides to DKP's; Applications of py-GC to sequence determination in ActlOomycins. Chemical Communications, 39-40.

't.'nn~ S. 191.)7. Bivmml'mliz<ltion: the form(ld),lhle p;ut of biolOorganic Chelnls!r},1 }tlllY/lIIl vf tilt' C/U'1nIC,lf Soell'tll-D.lltoll Tr.1JI5n:n.lI15, 21, 3953-3961". .

~ah,1r, S. & Tajmir-Riahi, H.A. 1995. Do metal ions alter the protein secondary structure of a light - harvesting com­pl,ex of. thylakoid membranes? IOl/mal of Inorganic Blocilellllstry, 58, 223-238.

Sekkal, M., Declerck, c., Huvenne, J-P., Legrand, P., Sombret, B. & Verdus M-e. 1993. Direct structural identification of polysaccharides from red algea by FTIR microspec­trometry II: Identification of the constituents of Gracl1aria verrl/cosa. Mikrochimica Acta, 112, 11-18.

Sidner, B.R., & McKee, T.R. 1976. Geochemical Controls on Vertical Distribution of Iron Rich Agglutinated For~minifers in Late Quaternary Continental Slope Sediments from Northwest Gulf of Mexico. AmericaJ! Association of Petroleum Geologists Bulletin, 60, 4, 722-733.

Simmonds, P.C. 1970. Whole microorganisms studied by Py­G.C.M.5; significance for E.T. life detection experiments. Applied Microbiology, 20, 4, 567-572.

Smith, KF., H.1f1S. P.L. Chapman, D., Reid, K.B.Y1. & Perkms, S.J. 19Y4, nl~ta-she('t s('('ond.lry structure uf trimeric glob-

ul:tr domain of Clq of complement "nd l'Ol1.1gen tvpes VIII and X .by FTIR spectroscopy and a\'l'r<lgl'~i ~truchlrl' oredlctlons. Blll(h.'IIIICJ/ lVI/mal. 301, 2·H-2:16

Stankiewicz, B.A., VanBergen, P.F., Duncan, I.J., Carter, J.F., Briggs, D.E.G. & Evershed, RP. 1996. Recognition of chitin and proteins in invertebrate cuticles using analyti­cal pxrolysis / Gas Chromatography and pyrolysis I Gas Chromatography / Mass Spectrometry. Rapid Communicatiolls ;'1 Mass Spectrometry, 10, ]747-]757.

Stankiewicz, A.B., Mastalerz, M., Kruge, M.A., Vanbergen, P.F. & Sadowska, A. 1997. A comparative study of modern and fossil cone scales and seeds of conifers a geochemical approach. New PIIyt%gy, 135, 375-393.

Stryer, L. 1975. Biocllemistry. pp. 11-70. Freeman. Tregelaar, E.W., Matthezing, R.M., Jansen, J.B.H., Horsfield,

B. & Deleeuw, J.W. 1989. Possible origins of normal­alkanes in high-wax crude oils. Nature, 342, 6429, 529-531

Viln def K:1"den, A, Boon, J.J. & dt..' Lel'uw, J.W lY84. Compari~H)n. of "nal~·b.cal pyrolysis tcchniques in the chilnctt.'rts.1Ih>O of chlttn. AnalytIcal c/u'/llIstrt/, 56, 2160· 2163 .

Whaley, W.G. 1975. Tile Golgi Apparatus. (Cell Biology MOllogrnplls, 2.) Springer, New York, 190 pp.

--....-<OO~o_--

Plate 1, As/rammina rara Rhumbler; 2. jadammina macreseens (Brady); 3. Trocilntllmina inftata Montagu; 4. Ammonia beeearii (Linne).